early-age shrinkage of ultra high-performance concrete
TRANSCRIPT
Western University Western University
Scholarship@Western Scholarship@Western
Electronic Thesis and Dissertation Repository
4-27-2011 12:00 AM
Early-Age Shrinkage of Ultra High-Performance Concrete: Early-Age Shrinkage of Ultra High-Performance Concrete:
Mitigation and Compensating Mechanisms Mitigation and Compensating Mechanisms
Ahmed Mohammed Soliman, The University of Western Ontario
Supervisor: Dr. Moncef L. Nehdi, The University of Western Ontario
A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree
in Civil and Environmental Engineering
© Ahmed Mohammed Soliman 2011
Follow this and additional works at: https://ir.lib.uwo.ca/etd
Part of the Other Civil and Environmental Engineering Commons, and the Structural Materials
Commons
Recommended Citation Recommended Citation Soliman, Ahmed Mohammed, "Early-Age Shrinkage of Ultra High-Performance Concrete: Mitigation and Compensating Mechanisms" (2011). Electronic Thesis and Dissertation Repository. 145. https://ir.lib.uwo.ca/etd/145
This Dissertation/Thesis is brought to you for free and open access by Scholarship@Western. It has been accepted for inclusion in Electronic Thesis and Dissertation Repository by an authorized administrator of Scholarship@Western. For more information, please contact [email protected].
EARLY-AGE SHRINKAGE OF ULTRA HIGH-PERFORMANCE CONCRETE: MITIGATION AND COMPENSATING MECHANISMS
(Spine title: Early-age shrinkage of ultra high-performance concrete)
(Thesis format: Integrated Article)
by
AHMED MOHAMMED SOLIMAN
Graduate Program in Engineering Science
Department of Civil and Environmental Engineering
A thesis submitted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
The School of Graduate and Postdoctoral Studies The University of Western Ontario
London, Ontario, Canada
© Ahmed Soliman 2011
THE UNIVERSITY OF WESTERN ONTARIO School of Graduate and Postdoctoral Studies
CERTIFICATE OF EXAMINATION
Supervisor ______________________________ Dr. Moncef L. Nehdi
Examiners ______________________________ Dr. Tim A. Newson ______________________________ Dr. Julie Q. Shang ______________________________ Dr. Robert J. Klassen ______________________________ Dr. Medhat Shetata
The thesis by
Ahmed Mohammed Soliman
entitled:
Early-Age Shrinkage of Ultra High-Performance Concrete: Mitigation and Compensating Mechanisms
is accepted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy
______________________ _______________________________ Date Chair of the Thesis Examination Board
ii
Abstract
The very high mechanical strength and enhanced durability of ultra high-
performance concrete (UHPC) make it a strong contender for several concrete
applications. However, UHPC has a very low water-to-cement ratio, which increases its
tendency to undergo early-age shrinkage cracking with a risk of decreasing its long-term
durability. To reduce the magnitude of early-age shrinkage and cracking
potential, several mitigation strategies have been proposed including the use of shrinkage
reducing admixtures, internal curing methods (e.g. superabsorbent polymers), expansive
cements and extended moist curing durations. To appropriately utilize these strategies, it
is important to have a complete understanding of the driving forces behind early-age
volume change and how these shrinkage mitigation methods work from a materials
science perspective to reduce shrinkage under field like conditions.
This dissertation initially uses a first-principles approach to understand the
interrelation mechanisms between different shrinkage types under simulated field
conditions and the role of different shrinkage mitigations methods. The ultimate goal of
the dissertation is to achieve lower early-age shrinkage and cracking risk concrete along
with reducing its environmental and economic impact. As a result, a novel
environmentally friendly shrinkage reducing technique based on using partially hydrated
cementitious materials (PHCM) from waste concrete is proposed. The PHCM principle,
mechanisms and efficiency were evaluated compared to other mitigation methods.
Furthermore, the potential of replacing cement with wollastonite microfibers was
investigated as a new strategy to produce UHPC with lower carbon foot-print, through
reducing the cement production. Finally, an artificial neural networks (ANN) model for
early-age autogenous shrinkage of concrete was proposed.
The evidence and insights provided by the experiments can be summarized in:
drying and autogenous shrinkage are dependant phenomena and applying the
conventional superposition principle will lead to an overestimation of the actual
autogenous shrinkage, adequately considering in-situ conditions in testing protocols
should allow gaining a better understanding of shrinkage mitigation mechanisms, the
iii
PHCM technique provides a passive internal restraining system that resists deformation
as early as the cementitious materials are mixed, wollastonite microfibers can act as an
internal restraint for shrinkage, reinforcing the microstructure at the micro-crack level
and leading to an enhancement of the early-age engineering properties, along with
gaining environmental benefits, and ANN showed success in predicting autogenous
shrinkage under simulated field conditions.
Keywords: Shrinkage, ultra high-performance concrete, field-like conditions,
superabsorbent polymer, shrinkage reducing admixture, wollastonite microfibers,
partially hydrated cementitious materials, drying/wetting cycles, carbon-oxygen demand
(COD), artificial neural networks.
iv
Co-Authorship Statement
This thesis has been prepared in accordance with the regulation of integrated-article
format stipulated by the Faculty of Graduate Studies at The University of Western
Ontario. Substantial parts of this thesis were either published in or submitted for
publication to peer-reviewed technical journals and an international conference. All
experimental work, data analysis, modeling process and writing of initial versions of all
publications listed below were carried out by the candidate himself. The contribution of
his research advisor and any other co-author, if applicable, consisted of either providing
advice, and/or helping in the development of the final versions of publications:
1. Nehdi, M. L. and Soliman, A. M. (Accepted 2009) " Early-age properties of concrete: Overview of fundamental concepts and state-of-the art research," Construction Materials, ICE., Vol. 164, Issue CM2, April 2011, pp. 55-77.
2. Soliman, A. M. and Nehdi, M. L. (2009) “Effect of drying conditions on autogenous shrinkage of ultra-high performance concrete at early-age," Materials and Structures, RILEM. In Press (published on line in 2011- DOI: 10.1617/s11527-010-9670-0).
3. Soliman, A. M. and Nehdi, M. L. (Accepted 2011) “Early-age shrinkage of ultra high performance concrete under drying/wetting cycles and submerged conditions," ACI Materials Journal, American Concrete Institute. Accepted.
4. Soliman, A. M. and Nehdi, M. L. (Accepted 2010) “Self-accelerated reactive powder concrete using partially hydrated cementitious materials," ACI Materials Journal, American Concrete Institute. In Press (to appear in 2011).
5. Soliman, A. M. and Nehdi, M. L. (2010) “Self-restraining concrete: mechanisms and Evidence," Cement and Concrete Research, Elsevier Science. Submitted.
6. Soliman, A. M. and Nehdi, M. L. (2010) “Can partially hydrated cementitious materials mitigate early-age shrinkage in ultra-high Performance concrete?,” Cement and Concrete Research, Elsevier Science. Submitted.
7. Soliman, A. M. and Nehdi, M. L. (2010) “Influence of natural wollastonite microfiber on early-age behaviour of ultra high-performance concrete," Journal of Materials in Civil Engineering, ASCE. Submitted.
8. Soliman, A. M. and Nehdi, M. L. (2011) “Shrinkage behaviour of ultra high performance concrete with shrinkage reducing admixture and wollastonite microfiber,” Cement and Concrete Composites, Elsevier Science. Submitted.
v
9. Soliman, A. M. and Nehdi, M. L. (2011) “Artificial neural network modeling of early-age autogenous shrinkage of concrete,” ACI Materials Journal, American Concrete Institute. Submitted.
10. Soliman, A. M. and Nehdi, M. L. (2010) “Performance of shrinkage reducing admixture under drying/wetting cycles and submerged conditions,” 2nd International Structures Specialty Conference, Winnipeg, Manitoba, Canada, (ST-30-1: ST-30-7). Published.
11. Soliman, A. M. and Nehdi, M. L. (2010) “Effect of Shrinkage Mitigation Methods on Early-Age Shrinkage of UHPC under simulated field conditions,” 8th International Conference on. Short and Medium Span Bridges, Niagara Falls, Ontario, Canada, 2010, pp. (83-1 :83-8) Published.
12. Soliman, A. M. and Nehdi, M. L. (2012) “Early-age behaviour of wollastonite microfiber UHPC,” Twelfth International Conference on Recent Advances in Concrete Technology and Sustainability Issues, Prague, Czech Republic, 2012. Submitted.
13. Soliman, A. M. and Nehdi, M. L. (2012) “Performance of concrete shrinkage mitigation admixtures under simulated field conditions,” Proceeding of the Tenth International Conference on Superplasticizer and Other Chemical Admixtures in Concrete to be held in Prague, Czech Republic, October 28 to 31, 2012. Submitted.
14. Soliman, A. M. and Nehdi, M. L. (2011), “Environmentally friendly self-accelerated concrete,” Proceedings of 2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted.
15. Soliman, A. M. and Nehdi, M. L. (2011), “Early-age behaviour of shrinkage reducing admixture and wollastonite microfibers in UHPC,” Proceedings of 2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted.
16. Soliman, A. M. and Nehdi, M. L. (2011), “Effect of exposure conditions on early-age shrinkage of UHPC thin elements,” Proceedings of 2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted.
vi
To: My Father Mohammed Salah El-Din,(God bless his soul)
My Mother Nemat,
My Sisters Rania, Yara, and Nanies
My Beloved Wife Noha,
My Beloved Daughter Malak,
And My Beloved Son Youssef.
vii
Acknowledgments
The author would like to express his sincere gratitude to his advisor, Prof. Moncef Nehdi,
for his guidance, advice and encouragement throughout the course of this work. His
mentorship, support and patience are highly appreciated. The author is indebted to the
University of Western Ontario for supporting his doctoral studies through the Western
Engineering Scholarship (WES). A grant provided by the Natural Science and
Engineering Research Council of Canada (NSERC) for manufacturing a state-of-the-art
walk-in environmental chamber used in the experimental research is highly appreciated.
Donations of materials by several companies (BASF, NYCO Minerals Inc., Emerging
Technologies inc. and Lafarge) were crucial to this research.
Special thanks are due to the undergraduate and graduate work-study students (Aiham
Adawi, Mohammed Mousa, and Emad Saleh) who assisted in the physically demanding
laboratory work. The author would also like to thank the University Machine Shop and
all Technicians and Staff at the Department of Civil and Environmental Engineering, who
contributed directly or indirectly to the accomplishment of this thesis.
Finally, the author would like to express his deep gratitude and appreciation to his father
(god bless his soul), mother, sisters, wife, beloved daughter and son for their continuous
support and encouragement throughout the course of this work. In particular, without my
wife's support and patience, this work would not have been possible.
viii
TABLE OF CONTENTS
CERTIFCATE OF EXAMINATION--------------------------------------------------- iiAbstract---------------------------------------------------------------------------------------- iiiCo-Authorship Statement------------------------------------------------------------------ vAcknowledgments---------------------------------------------------------------------------- viiiTable of Contents---------------------------------------------------------------------------- ixList of Tables---------------------------------------------------------------------------------- xviiList of Figures-------------------------------------------------------------------------------- xixList of Appendices------------------------------------------------------------------------- xxviii
CHAPTER ONE
INTRODUCTION
1.1. ULTRA HIGH PERFORMANCE CONCRETE (UHPC)------------------------ 11.2. EARLY-AGE SHRINKAGE--------------------------------------------------------- 31.3. NEED FOR RESEARCH-------------------------------------------------------------- 51.4. OBJECTIVES AND SCOPE---------------------------------------------------------- 61.5. STRUCTURE OF THESIS------------------------------------------------------------ 81.6. ORIGINAL CONTRIBUTIONS----------------------------------------------------- 91.7. REFERENCES-------------------------------------------------------------------------- 13
CHAPTER TWO
EARLY-AGE PROPERTIES OF CONCRETE: OVERVIEW
2.1. INTRODUCTION---------------------------------------------------------------------- 162.2. EARLY-AGE PROPERTIES--------------------------------------------------------- 18
2.2.1 Early-Age Thermal Properties---------------------------------------------------- 182.2.1.1 Heat of Hydration--------------------------------------------------------------- 192.2.1.2 Specific Heat Capacity--------------------------------------------------------- 222.2.1.3 Thermal Conductivity---------------------------------------------------------- 232.2.1.4 Thermal Diffusivity------------------------------------------------------------- 252.2.1.5 Coefficient of Thermal Expansion-------------------------------------------- 27
2.2.2 Early-Age Mechanical Properties-------------------------------------------------- 30
ix
2.2.2.1 Compressive Strength---------------------------------------------------------- 302.2.2.2 Tensile Strength---------------------------------------------------------------- 332.2.2.3 Modulus of Elasticity---------------------------------------------------------- 392.2.2.4 Poisson’s Ratio----------------------------------------------------------------- 42
2.2.3. Early-Age Shrinkage of Concrete------------------------------------------------- 432.2.3.1 Thermal Dilation--------------------------------------------------------------- 442.2.3.2 Drying Shrinkage--------------------------------------------------------------- 452.2.3.3 Autogenous Shrinkage--------------------------------------------------------- 462.2.3.4 Factors Affecting Early-Age Shrinkage------------------------------------- 502.2.3.5 Compensating for Early-Age Shrinkage------------------------------------ 54
2.2.4. Early-Age Creep of Concrete----------------------------------------------------- 552.3. MODELS FOR SIMULATION AND ANALYSIS OF EARLY-AGE
PROPERTIES---------------------------------------------------------------------------- 592.4. CONCLUDING REMARKS---------------------------------------------------------- 612.5. REFERENCES-------------------------------------------------------------------------- 63
CHAPTER THREE
INTERACTIONS MECHANSIMS OF EARLY-AGE SHRINKAGE OF UHPC UNDER SIMULATED DRYING CONDITIONS
3.1. INTRODUCTION----------------------------------------------------------------------- 733.2. RESARCH SIGNIFICANCE---------------------------------------------------------- 753.3. METHODOLOGY---------------------------------------------------------------------- 753.4. EXPERIMENTAL PROGRAM------------------------------------------------------- 76
3.4.1. Materials and Mixture Proportions----------------------------------------------- 763.4.2. Environmental Conditions-------------------------------------------------------- 763.4.3. Preparation of Test Specimens and Testing Procedures----------------------- 78
3.4.3.1 Chemically Bound Water and Degree of Hydration------------------------ 793.4.3.2 Strain Measurements------------------------------------------------------------ 803.4.3.3 Moisture Loss-------------------------------------------------------------------- 813.4.3.4 Coefficient of Thermal Expansion-------------------------------------------- 81
3.5. RESULTS AND DISCUSSION------------------------------------------------------ 833.5.1. Compressive Strength--------------------------------------------------------------- 833.5.2. Total Strain--------------------------------------------------------------------------- 86
3.5.2.1 Influence of Curing Temperature and Ambient Humidity on Total Strain----------------------------------------------------------------------------- 86
x
3.5.3. Autogenous Strain------------------------------------------------------------------ 873.5.3.1 Effect of Curing Temperature and w/c-------------------------------------- 87
3.5.4. Degree of Hydration--------------------------------------------------------------- 913.5.4.1 Under Sealed Condition------------------------------------------------------- 913.5.4.2 Under Drying Condition------------------------------------------------------- 93
3.5.5. Correlation between Autogenous Strain and Achieved Degree of Hydration under Sealed Condition----------------------------------------------- 98
3.5.6. Predicting Drying and Autogenous Strains under Different Curing Conditions--------------------------------------------------------------------------- 99
3.5.6.1 Superposition Principle-------------------------------------------------------- 993.5.6.2 Degree of Hydration Method------------------------------------------------- 101
3.6. CONCLUSIONS ----------------------------------------------------------------------- 1043.7. REFERENCES-------------------------------------------------------------------------- 105
CHAPTER FOUR
INFLUENCE OF SHRINKAGE MITIGATION TECHNIQUES ON EARLY-AGE SHRINKAGE INTERRELATION UNDER DRYING CONDITIONS
4.1. INTRODUCTION----------------------------------------------------------------------- 1104.2. RESEARCH SIGNIFICANCE-------------------------------------------------------- 1114.3. METHODOLOGY---------------------------------------------------------------------- 1114.4. EXPERIMENTAL PROGRAM ------------------------------------------------------- 112
4.4.1. Materials and Mixture Proportions----------------------------------------------- 1124.4.2. Environmental Conditions-------------------------------------------------------- 1124.4.3. Preparation of Test Specimens and Testing Procedures---------------------- 113
4.5. RESULTS AND DISCUSSION------------------------------------------------------ 1134.5.1. Compressive Strength-------------------------------------------------------------- 1134.5.2. Total Strain-------------------------------------------------------------------------- 1154.5.3. Autogenous Strain------------------------------------------------------------------ 116
4.5.3.1 Effect of Curing Temperatures----------------------------------------------- 1164.5.3.2 Effect of w/c under Different Curing Temperatures---------------------- 119
4.5.4. Degree of Hydration--------------------------------------------------------------- 1224.5.4.1 Under Sealed Condition ------------------------------------------------------- 1224.5.4.2 Under Drying Condition------------------------------------------------------- 123
xi
4.5.4.2.1 SRA mixtures-------------------------------------------------------------- 1234.5.4.2.2 SAP mixtures-------------------------------------------------------------- 128
4.5.5. Correlation between Autogenous Strain and Achieved Degree of Hydration under Sealed Condition----------------------------------------------- 132
4.5.5.1 SRA Mixtures------------------------------------------------------------------- 1324.5.5.2 SAP Mixtures------------------------------------------------------------------- 133
4.5.6. Predicting Drying and Autogenous Strains under Different Curing Conditions--------------------------------------------------------------------------- 134
4.5.6.1 SRA Mixtures------------------------------------------------------------------- 1354.5.6.2 SAP Mixtures------------------------------------------------------------------- 137
4.6. CONCLUSIONS ------------------------------------------------------------------------ 1394.7. REFERENCES--------------------------------------------------------------------------- 141
CHAPTER FIVE
EARLY-AGE SHRINKAGE OF UHPC UNDER DRYING/WETTING CYCLES AND SUBMERGED CONDITIONS
5.1. INTRODUCTION---------------------------------------------------------------------- 1445.2. RESEARCH SIGNIFICANCE------------------------------------------------------- 1465.3. EXPERIMENTAL PROCEDURE--------------------------------------------------- 146
5.3.1. Materials and Mixture Proportions---------------------------------------------- 1465.3.2. Environmental Conditions-------------------------------------------------------- 1465.3.3. Preparation of Test Specimens and Testing Procedures---------------------- 147
5.3.3.1 Chemically Bound Water and Degree of Hydration---------------------- 1485.3.3.2 Strain Measurements---------------------------------------------------------- 1485.3.3.3 Moisture Loss------------------------------------------------------------------ 1505.3.3.4 Mercury Intrusion Porosimetry (MIP)-------------------------------------- 1505.3.3.5 Chemical Oxygen Demand (COD)----------------------------------------- 150
5.4. RESULTS AND DISCUSSION----------------------------------------------------- 1515.4.1. Mass Change----------------------------------------------------------------------- 1515.4.2. Shrinkage Strain under Different Environmental Exposure Conditions--- 155
5.4.2.1 Control Mixture---------------------------------------------------------------- 1555.4.2.2 Effect of SRA------------------------------------------------------------------- 1605.4.2.3 Effect of SAP------------------------------------------------------------------- 1655.4.2.4 Synergistic Effect of SRA and SAP----------------------------------------- 171
xii
5.5. CONCLUSIONS------------------------------------------------------------------------- 1735.6. REFERENCES--------------------------------------------------------------------------- 174
CHAPTER SIX
SELF-RESTRAINING SHRINKAGE CONCRETE: MECHANISMS AND EVIDENCE
6.1. INTRODUCTION---------------------------------------------------------------------- 1786.2. EARLY-AGE PHASES AND AUTOGENOUS SHRINKAGE IN CEMENTITIOUS MATERIALS---------------------------------------------------- 1806.3. EARLY-AGE SHRINKAGE MITIGATION TECHNIQUES------------------- 1816.4. SHRINKAGE RESTRAINING SYSTEMS---------------------------------------- 1856.5. CAN PHCM ADDITION REDUCE CONCRETE SHRINKAGE-------------- 1866.6. SOURCES OF PHCM----------------------------------------------------------------- 1906.7. PARTIALLY HYDRATED CEMENTITIOUS MATERIALS METHODOLOGY--------------------------------------------------------------------- 1916.8. EXPERIMENTAL PROCEDURE--------------------------------------------------- 191
6.8.1. Materials and Mixture Proportions---------------------------------------------- 1926.8.2. Preparation of Test Specimens and Testing Procedures---------------------- 193
6.9. RESULTS AND DISCUSSION------------------------------------------------------ 1966.9.1. Setting Time and Compressive Strength--------------------------------------- 1966.9.2. Degree of Hydration-------------------------------------------------------------- 2006.9.3. Heat of Hydration----------------------------------------------------------------- 2006.9.4. Shrinkage--------------------------------------------------------------------------- 202
6.10. MECHANISM OF PHCM ACTION---------------------------------------------- 2056.11. CONCLUSIONS---------------------------------------------------------------------- 2116.12. REFERENCES------------------------------------------------------------------------ 212
CHAPTER SEVEN
EVALUATION OF SELF-RESTRAINING SHRINKAGE TECHNIQUE COMPARED TO CONVENTIONAL METHODS IN UHPC: INDIVDUALLY AND COMBINED
7.1. INTRODUCTION--------------------------------------------------------------------- 2177.2. RESEARCH SIGNIFICANCE------------------------------------------------------ 219
xiii
7.3. EXPERIMENTAL PROGRAM----------------------------------------------------- 2197.3.1. Materials and Mixture Proportions--------------------------------------------- 2207.3.2. Preparation of Test Specimens and Testing Procedures--------------------- 221
7.4. RESULTS AND DISCUSSION----------------------------------------------------- 2217.4.1. Setting Time and Early-Age Compressive Strength------------------------- 2217.4.2. Degree of Hydration-------------------------------------------------------------- 2247.4.3. Heat of Hydration----------------------------------------------------------------- 2287.4.4. Shrinkage and Mass Loss-------------------------------------------------------- 231
7.5. STATISTICAL ANALYSIS FOR EFFECT OF PHCM ADDITION ON SHRINKAGE-------------------------------------------------------------------------- 2437.6. CONCLUSIONS---------------------------------------------------------------------- 2457.7. REFERENCES------------------------------------------------------------------------- 246
CHAPTER EIGHT
INFLUENCE OF NATURAL WOLLASTONITE MICROFIBERS ON EARLY-AGE BEHAVIOUR OF UHPC
8.1. INTRODUCTION-------------------------------------------------------------------- 2508.2. RESEARCH SIGNIFICANCE------------------------------------------------------ 2528.3. EXPERIMENTAL PROGRAM ---------------------------------------------------- 253
8.3.1. Materials and Mixture Proportions-------------------------------------------- 2538.3.2. Preparation of Test Specimens and Testing Procedures-------------------- 254
8.4. RESULTS AND DISCUSSION ----------------------------------------------------- 2568.4.1. Workability ----------------------------------------------------------------------- 2568.4.2. Early-Age Compressive Strength----------------------------------------------- 2578.4.3. Heat of Hydration----------------------------------------------------------------- 2628.4.4. Free Shrinkage and Mass Loss-------------------------------------------------- 2668.4.5. Restrained Shrinkage------------------------------------------------------------- 2728.4.6. Flexural Strength------------------------------------------------------------------ 275
8.6. CONCLUSIONS---------------------------------------------------------------------- 2788.7. REFERENCES------------------------------------------------------------------------- 280
CHAPTER NINE
SHRINKAGE BEHAVIOUR OF UHPC WITH SHRINKAGE REDUCING ADMIXTURE AND WOLLASTONITE MICROFIBERS
9.1. INTRODUCTION--------------------------------------------------------------------- 285
xiv
9.2. RESEARCH SIGNIFICANCE------------------------------------------------------ 2869.3. EXPERIMENTAL PROGRAM----------------------------------------------------- 287
9.3.1. Materials and Mixture Proportions--------------------------------------------- 2879.3.2. Preparation of Test Specimens and Testing Procedures--------------------- 288
9.4. RESULTS AND DISCUSSION----------------------------------------------------- 2899.4.1. Compressive Strength------------------------------------------------------------ 2899.4.2. Heat of Hydration----------------------------------------------------------------- 2929.4.3. Degree of Hydration-------------------------------------------------------------- 2959.4.4. Free Shrinkage and Mass Loss-------------------------------------------------- 2969.4.5. Restrained Shrinkage------------------------------------------------------------- 3019.4.6. COD Results----------------------------------------------------------------------- 304
9.6. CONCLUSIONS---------------------------------------------------------------------- 3089.7. REFERENCES------------------------------------------------------------------------- 309
CHAPTER TEN
ARTIFICIAL NEURAL NETWORK MODELING OF EARLY-AGE AUTOGENOUS SHRINKAGE OF CONCRETE
10.1. INTRODUCTION------------------------------------------------------------------- 31210.2. RESEARCH SIGNIFICANCE----------------------------------------------------- 31510.3. NEURAL NETWORK APPROACH--------------------------------------------- 315
10.3.1. Feed-Forward Neural Network----------------------------------------------- 31710.3.2. Back-Propagation Learning Algorithm------------------------------------- 318
10.4. PROPOSED ANN MODEL-------------------------------------------------------- 31910.5. MODEL DATABASE-------------------------------------------------------------- 32210.6. RESULTS AND DISCUSSION--------------------------------------------------- 323
10.6.1. Validating ANN Using Experimental Work-------------------------------- 32610.6.2. Parametric Analyses----------------------------------------------------------- 329
10.6.2.1 Effect of Water-to-Cement Ratio--------------------------------------- 33010.6.2.2 Effect of Superplasticizer------------------------------------------------ 33210.6.2.3 Effect of Curing Temperature------------------------------------------- 334
10.7. CONCLUSIONS--------------------------------------------------------------------- 33510.8. REFERENCES----------------------------------------------------------------------- 337
xv
CHAPTER ELEVEN
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
11.1. SUMMARY AND CONCLUSIONS----------------------------------------------- 34311.2. RECOMMENDATION FOR FUTURE WORK---------------------------------- 351 APPENDICES------------------------------------------------------------------------------- 354VITA------------------------------------------------------------------------------------------ 377
xvi
LIST OF TABLES
Table 2- 1: Definition of early-age period -----------------------------------------------
18
Table 2- 2: Heat of hydration of cement compounds [after (McCullough and Rasmussen, 1999), (Lea, 2004)]---------------------------------------------
20
Table 2- 3: Values of specific heat investigated by different researchers ----------- 23
Table 2- 4: Thermal properties of concrete with different aggregate types [after (Mehta and Monteiro, 2006)] -----------------------------------------------
25
Table 2- 5: Values of thermal diffusivity obtained by different researchers-------- 27
Table 2- 6: Values of coefficient of thermal expansion obtained by different researchers---------------------------------------------------------------------
29
Table 2- 7: Different suggested values for k and n------------------------------------- 36
Table 2- 8: Modulus of elasticity as a function of compressive strength of concrete at any age------------------------------------------------------------
41
Table 2- 9: Poisson’s ratio values--------------------------------------------------------- 43
Table 2- 10: Methods used for measuring shrinkage at early-age---------------------- 49
Table 2- 11: Shrinkage compensating methods------------------------------------------- 55
Table 3- 1: Chemical and physical properties of cement and silica fume----------- 77
Table 3- 2: Composition of control mixture--------------------------------------------- 78
Table 3- 3: Simulated environmental conditions---------------------------------------- 78
Table 3- 4: Overestimation ratios for autogenous strains under different curing conditions-----------------------------------------------------------------------
102
Table 4- 1: Overestimation ratios for autogenous strains under different curing conditions for different w/c=0.25 UHPC mixtures-----------------------
137
Table 5- 1: Computation of (rs) for control and SAP mixtures under drying conditions and drying/wetting cycles---------------------------------------
168
xvii
Table 5- 2: Percentage of reduction in shrinkage strain for each exposure condition due to incorporating both SRA and SAP compared to reference mixtures-------------------------------------------------------------
172
Table 6- 1: Composition of control and PHCM mixtures------------------------------ 192
Table 6- 2: Tested mixtures---------------------------------------------------------------- 192
Table 6- 3: Initial and final setting times for different mixtures---------------------- 198
Table 7- 1: Tested mixtures---------------------------------------------------------------- 220
Table 7- 2: Setting time results for the tested mixtures-------------------------------- 224
Table 8- 1: Tested Mixtures---------------------------------------------------------------- 254
Table 8- 2: Compressive strength results of UHPC mixtures incorporating different content and sizes of wollastonite microfibers------------------
258
Table 9- 1: Tested mixtures---------------------------------------------------------------- 288
Table 10- 1: The values of parameters used in the ANN model------------------------ 321
Table 10- 2: Database sources and variables range of input and output--------------- 323
Table B- 1: Tested mixtures---------------------------------------------------------------- 356
Table B- 2: Initial and final setting times for different mixtures----------------------- 358
Table C- 1: Sample for data used in ANN Model---------------------------------------- 373
Table E-1: Trial mixtures composition--------------------------------------------------- 377
xviii
LIST OF FIGURES
Figure 2-1: Heat of hydration developed after 72 hours at different temperatures (Type I: Ordinary portland cement, Type II: Moderate sulphate resistance portland cement, Type IV: Low heat portland cement) [adapted after (Neville, 1996)]----------------------------------------------
21
Figure 2-2: Evaluation of the coefficient of thermal expansion of concrete during hydration [adapted after (Kada et al., 2002)]---------------------
28
Figure 2-3: Average additional maturity, (F-hr), due to hydration [adapted after (Oluokun et al., 1990)]--- ---------------------------------------------------
33
Figure 2-4: Tensile / compressive strength of concrete versus age------------------ 35
Figure 2-5: a) Uniaxial tensile strength, and b) Uniaxial tensile strain capacity [adapted after (Hammer et al., 2007)]-------------------------------------
37
Figure 2-6: Tensile strength of concrete with different water/cement ratio [after (Abel and Hover, 1998)]-----------------------------------------------------
38
Figure 2-7: Tensile strength of concrete with shrinkage reducing admixture [adapted after (D’Ambrosia, 2002)]----------------------------------------
39
Figure 2-8: Modulus of elasticity to compressive strength ratio data classified by concrete age [after Zhao, 1990]---------------------------------------------
41
Figure 2-9: Accumulation of early-age and long-term shrinkage, with various curing environments (Wind = 2 m/s (4.5 mph), dry = 40% RH, wet = 100% RH) [after (Holt, 2005)]-------------------------------------------
44
Figure 2-10: Chemical shrinkage and autogenous shrinkage of concrete [adapted after Mihashi and Leite (2004)]---------------------------------------------
47
Figure 2-11: Autogenous and chemical shrinkage during different stages, as a function of degree of hydration [adapted after (Holt, 2001, Esping, 2007)]--------------------------------------------------------------------------
48
Figure 2-12: Free shrinkage with different w/c and 1% shrinkage reducing admixture [adapted after (D’Ambrosia, 2002)]---------------------------
53
Figure 2-13: Apparent and real creep compliance function [after (Lee et al., 2006)]--------------------------------------------------------------------------
58
Figure 3-1: Coefficient of thermal expansion test setup------------------------------- 83
xix
Figure 3-2: Effect of curing temperature on compressive strength for mixtures with w/c=0.25-----------------------------------------------------------------
84
Figure 3-3: Effect of ambient humidity on compressive strength for mixtures with w/c=0.25-----------------------------------------------------------------
85
Figure 3-4: Effect of w/c on compressive strength------------------------------------- 86
Figure 3-5: Total strains Vs. ambient humidity at different curing temperatures for a) w/c=0.25 and b) w/c=0.22 mixtures at age of 7-days------------
88
Figure 3-6: Autogenous strains for the control mixture (w/c=0.25) at 10, 20 and 40°C----------------------------------------------------------------------------
89
Figure 3-7: Autogenous strains for mixtures with w/c=0.22 and 0.25 at different temperatures-------------------------------------------------------------------
90
Figure 3-8: Scanning electron micrograph for w/c =0.22 UHPC specimens: (a) agglomerated silica fume encapsulated with C-S-H, (b) elemental distribution---------------------------------------------------------------------
91
Figure 3-9: Evaluation of degree of hydration Vs. age for control mixtures a) w/c=0.25 and b) w/c= 0.22-------------------------------------------------
92
Figure 3-10: Cross sections for a) sealed and b) exposed to drying specimens after 7 days--------------------------------------------------------------------
94
Figure 3-11: Change in BW Vs. ambient humidity for w/c=0.25 control mixtures at a) 10°C, b) 20°C, and c) 40°C--------------------------------------------
95
Figure 3-12: Change in mass loss Vs. ambient humidity for w/c=0.25 control mixtures at a) 10°C, b) 20°C, and c) 40°C---------------------------------
97
Figure 3-13: Correlation between autogenous strains and RBW for control mixtures cured at 10, 20 and 40°C-----------------------------------------
99
Figure 3-14: Drying strains for w/c=0.25 control mixtures based on the superposition principle at age of 7-days-----------------------------------
100
Figure 3-15: Autogenous strains evaluated based on achieved degree of hydration for w/c=0.25 UHPC mixtures at age 7 days------------------------------
102
Figure 3-16: Drying strains evaluated based on achieved degree of hydration for w/c=0.25 UHPC mixtures at age 7 days-----------------------------------
103
xx
Figure 4-1: Effect of adding 2% SRA and 0.6%SAP on compressive strength after 7 days for mixtures with w/c=0.25-----------------------------------
114
Figure 4-2: Total strains Vs. ambient humidity at different curing temperatures for w/c=0.25 mixtures at age of 7-days: a) SRA, and b) SAP----------
115
Figure 4-3: Autogenous strains after 7 days for mixtures with w/c=0.25----------- 117
Figure 4-4: Autogenous strains at 10°C: a) observed, and b) after expansion for mixtures with w/c=0.25------------------------------------------------------
118
Figure 4-5: Scanning electron micrograph for agglomerated silica fume in SRA mixture with w/c=0.22-------------------------------------------------------
120
Figure 4-6: Autogenous strains for mixtures with w/c=0.22 and 0.25 at different temperatures a) control, b)SRA and c) SAP mixtures-------------------
121
Figure 4-7: Evaluation of degree of hydration Vs. age for a) w/c=0.25 + 2% SRA, and b) w/c=0.25 + 0.06% SAP UHPC mixtures cured at 10, 20 and 40°C-------------------------------------------------------------------
122
Figure 4-8: Change in BW Vs. ambient humidity for w/c=0.25 + 2% SRA mixtures at a) 10°C, b) 20°C, and c) 40°C--------------------------------
124
Figure 4-9: Cross section for SRA specimen exposed to drying condition after 7 days-----------------------------------------------------------------------------
126
Figure 4-10: Change in mass loss Vs. ambient humidity for w/c=0.25 + 2% SRA mixtures at a) 10°C, b) 20°C, and c) 40°C--------------------------------
127
Figure 4-11: Change in BW Vs. ambient humidity for w/c=0.25 + 0.6% SAP mixtures at a) 10°C, b) 20°C, and c) 40°C--------------------------------
129
Figure 4-12: Change in mass loss Vs. ambient humidity for w/c=0.25 + 0.6% SAP mixtures at a) 10°C, b) 20°C, and c) 40°C--------------------------
130
Figure 4-13: Correlation between autogenous strains and RBW for 2% SRA mixtures cured at 10, 20 and 40°C-----------------------------------------
133
Figure 4-14: Correlation between autogenous strains and RBW for 0.6% SAP mixtures cured at 10, 20 and 40°C-----------------------------------------
134
Figure 4-15: Autogenous strains evaluated based on achieved degree of hydration for w/c=0.25 + 2% SRA UHPC mixtures at age 7 days-----------------
136
xxi
Figure 4-16: Comparison between autogenous strains evaluated based on achieved degree of hydration for w/c=0.25 mixtures at age of 7-days with and without SRA-------------------------------------------------------
136
Figure 4-17: Autogenous strains evaluated based on achieved degree of hydration for w/c=0.25 + 0.6% SAP UHPC mixtures at age 7 days-------------
138
Figure 4-18: Comparison between autogenous strains evaluated based on achieved degree of hydration for w/c=0.25 mixtures at age of 7-days with and without SAP--------------------------------------------------------
139
Figure 5-1: (a) Strain measurement specimen, and (b) strain gauges arrangement to form full bridge circuit----------------------------------------------------
149
Figure 5-2: Mass change of control mixtures a) w/c=0.22 and b) w/c=0.25 under different exposure conditions-----------------------------------------------
152
Figure 5-3: Mass change for a) w/c=0.22 and b) w/c= 0.25 mixtures incorporating 2% SRA or 0.6% SAP under different exposure conditions----------------------------------------------------------------------
154
Figure 5-4: Measured strain for w/c=0.25 control specimens under different exposure conditions----------------------------------------------------------
155
Figure 5-5: Measured strain for w/c=0.25 control specimens under drying/ wetting cycles-----------------------------------------------------------------
156
Figure 5-6: Measured (a) porosity and (b) degree of hydration for w/c=0.25 control specimens surface (S) and core (C) under different exposure conditions----------------------------------------------------------------------
157
Figure 5-7: Pore size distribution for control specimen’s surface under different exposure conditions after the first drying/wetting cycles---------------
160
Figure 5-8: Measured strain for w/c=0.25 UHPC specimens incorporating 2% SRA under different exposure conditions---------------------------------
161
Figure 5-9: Drying strain rate during drying/wetting cycles for control and SRA UHPC specimens-------------------------------------------------------------
162
Figure 5-10: Measured strain after the early expansion period under submerged condition-----------------------------------------------------------------------
163
Figure 5-11: COD values for control and SRA UHPC specimens under (a)submerged condition and (b) drying/wetting cycles------------------
164
xxii
Figure 5-12: Measured strain for w/c=0.25 UHPC specimens incorporating 0.06% SAP under different exposure conditions-------------------------
166
Figure 5-13: Pore size percentage for w/c=0.25 control and SAP UHPC specimens under drying conditions----------------------------------------
167
Figure 5-14: Degree of hydration for w/c=0.25 control and SAP UHPC specimens under drying/wetting cycles------------------------------------
169
Figure 5-15: Measured strain for control and SAP UHPC specimens under SM condition after early expansion---------------------------------------------
171
Figure 6-1: Structure development and volume change during different early-age stages----------------------------------------------------------------------
183
Figure 6-2: Micro-crystals role in restraining deformation--------------------------- 186
Figure 6-3: NMR spectra for a) the anhydrous Portland cement, and b) SF-------- 193
Figure 6-4: The corrugated tube protocol a) test setup, and b) measuring unit----- 195
Figure 6-5: Early-age compressive strength of UHPC mixtures incorporating PHCM cured at a) 10°C and b) 20°C--------------------------------------
197
Figure 6-6: 28 days compressive strength of UHPC mixtures incorporating PHCM cured at 10°C and 20°C---------------------------------------------
199
Figure 6-7: Relation between degree of hydration (amount of BW) and compressive strength development for the PHCM mixtures-----------
200
Figure 6-8: Heat of hydration for UHPC mixtures incorporating PHCM--------- - 202
Figure 6-9: Total shrinkage for PHCM mixtures--------------------------------------- 203
Figure 6-10: Mass loss for PHCM mixtures---------------------------------------------- 203
Figure 6-11: Autogenous shrinkage for PHCM mixtures------------------------------- 204
Figure 6-12: NMR spectra for a) C and b) H2 samples at different ages------------- 206
Figure 6-13: Enthalpy values for C and H2 mixtures during the first 24 hrs--------- 208
Figure 6-14: Microstructure of H2 mixtures after a) 8, and b) 24 hrs as in backscattered electron microscopy, and energy dispersive X-ray element analysis--------------------------------------------------------------
210
xxiii
Figure 7-1: Early-age compressive strength development for mixtures incorporating a)SRA and/or SAP, and b)combined shrinkage mitigation techniques--------------------------------------------------------
223
Figure 7-2: Degree of hydration development for mixtures incorporating a)single and b) combined shrinkage mitigation techniques-------------
227
Figure 7-3: Heat of hydration development for mixtures incorporating a)SRA and/or SAP, and b)combined shrinkage mitigation techniques --------
230
Figure 7-4: Mass loss for mixtures incorporating a)single and b) combined shrinkage mitigation techniques -------------------------------------------
231
Figure 7-5: Shrinkage development for mixtures incorporating a)single and b) combined shrinkage mitigation techniques -------------------------------
232
Figure 7-6: Mass loss-shrinkage relationship for mixtures incorporating a)single and b)combined shrinkage mitigation techniques -----------------------
234
Figure 7-7: Autogenous shrinkage for mixtures incorporating a)single and b) combined shrinkage mitigation techniques -------------------------------
237
Figure 7-8: TGA curves at different ages for PHCM mixture with and without SRA-----------------------------------------------------------------------------
240
Figure 7-9: Thermal analysis different ages for PHCM mixture with and without SAP a) TGA and b) CH content--------------------------------------------
241
Figure 8-1: Different sizes of wollastonite microfibers a) MF1, b) MF2 and c)MF3--------------------------------------------------------------------------
254
Figure 8-2: Concrete ring test------------------------------------------------------------- 256
Figure 8-3: Relative change in flowability index for mixtures incorporating different content and sizes of wollastonite microfibers compared to that of the control mixture---------------------------------------------------
257
Figure 8-4: Very early-age compressive strength development for different wollastonite microfibers a) MF1, b) MF2 and c) MF3------------------
259
Figure 8-5: Relative early-age compressive strength of mixtures incorporating different content and sizes of wollastonite microfibers at age 7 days-
261
xxiv
Figure 8-6: a) General trend of heat of hydration for wollastonite microfibers mixtures and b) variation in temperature peak of wollastonite mixtures Vs. control mixture -----------------------------------------------
263
Figure 8-7: a) Development of degree of hydration and b) relative porosity for mixture incorporating 8% of wollastonite microfibers------------------
265
Figure 8-8: Mass loss for mixtures incorporating wollastonite microfibers with a) 4%, b) 8% and c) 12% contents compared to that of the control mixture-------------------------------------------------------------------------
266
Figure 8-9: a) General trend of total shrinkage for wollastonite microfibers mixtures and b) variation in total shrinkage of wollastonite mixtures Vs. control mixture at 7-days-----------------------------------------------
269
Figure 8-10: a) General trend of autogenous shrinkage for wollastonite microfibers mixtures and b) variation in autogenous shrinkage of wollastonite mixtures Vs. control mixture at 7-days --------------------
271
Figure 8-11: Strain measurements of the steel ring for mixtures incorporating different content of a) MF1 and b) MF3-----------------------------------
273
Figure 8-12: Total crack width for mixtures incorporating different content and sizes of wollastonite microfibers-------------------------------------------
275
Figure 8-13: Relative flexural strength for mixtures incorporating different content and sizes of wollastonite microfibers compared to that of control mixture----------------------------------------------------------------
276
Figure 8-14: Scanning electron micrograph and energy dispersive X-ray element analysis for ruptured wollastonite microfibers----------------------------
278
Figure 9- 1: Development of early-age compressive strength of UHPC mixtures with and without different dosages of SRA or wollastonite microfibers---------------------------------------------------------------------
289
Figure 9- 2: Effect of combing SRA and wollastonite microfibers on early-age compressive strength of UHPC with respect to a) SRA and b) wollastonite microfibers mixtures------------------------------------------
291
Figure 9- 3: Variation in the compressive strength of UHPC mixture with different dosages of SRA and/or wollastonite microfibers compared to the control mixture--------------------------------------------------------
292
Figure 9- 4: Heat of hydration for UHPC mixtures incorporating SRA and/or wollastonite microfibers a) individually and b) combined--------------
294
xxv
Figure 9- 5: Change in degree of hydration during the first day for different UHPC mixtures---------------------------------------------------------------
295
Figure 9- 6: Mass loss for UHPC mixtures with and without different dosages of SRA and/or wollastonite microfibers--------------------------------------
297
Figure 9- 7: Shrinkage strains for UHPC mixtures with and without different dosages of SRA and/or wollastonite microfibers under a) drying and b) sealed conditions----------------------------------------------------------
299
Figure 9- 8: SEM for wollastonite microfibers pullout from cementitious matrix incorporating SRA------------------------------------------------------------
301
Figure 9- 9: Strain measurements of the steel ring for UHPC mixtures incorporating SRA and/or wollastonite microfibers a) individually and b) combined--------------------------------------------------------------
303
Figure 9- 10: Measured strains under submerged condition for control, SRA and/or wollastonite microfibers UHPC mixtures: a) observed and b) initiated to zero after the early expansion---------------------------------
305
Figure 9- 11: COD values for control, SRA and/or wollastonite microfibers UHPC mixtures------------------------------------------------------------------------
307
Figure 10- 1: The architecture of ANN models------------------------------------------- 317
Figure 10- 2: Response of ANN model in predicting autogenous shrinkage of concrete------------------------------------------------------------------------
326
Figure 10- 3: Measured autogenous shrinkage values compared with predicted a) w/c=0.22 and b) w/c=0.25---------------------------------------------------
327
Figure 10- 4: Deviation between measured and predicted autogenous shrinkage values---------------------------------------------------------------------------
328
Figure 10- 5: Residual values for ANN model-------------------------------------------- 329
Figure 10- 6: Sensitivity of ANN model to the w/c in predicting autogenous shrinkage-----------------------------------------------------------------------
331
Figure 10- 7: Correlation between autogenous shrinkage of concrete measured at T=20°C and w/c at ages 24 and 72 hours---------------------------------
332
Figure 10- 8: Effect of superplasticizer dosage on autogenous shrinkage of concrete measured at T=20°C and w/c=0.26-----------------------------
334
xxvi
Figure 10- 9: Autogenous shrinkage vs. age for concrete w/c=0.26 and cured at various temperature ranges from 10 to 40°C-----------------------------
335
Figure A-1: Detailed temperature and total strain measured in concrete samples over 24 hours------------------------------------------------------------------
354
Figure A-2 Time-evolution of coefficient of thermal expansion--------------------- 354
Figure B-1 Early-age compressive strength of UHPC mixtures incorporating PHCM cured at a) 10°C and b) 20°C and UHPC mixtures incorporating CA and NCA cured at c) 10°C and d) 20°C-------------
357
Figure B-2 Compressive strength difference between UHPC mixtures incorporating PHCM and that incorporating CA [ a) MAC1, and b) MAC2] and NCA [c) MANC1, and d) MANC2] at different curing temperatures-------------------------------------------------------------------
361
Figure B-3 28 days compressive strength of UHPC mixtures incorporating PHCM , CA and NCA cured at 10°C and 20°C--------------------------
362
Figure B-4 Change in initial and final setting times for UHPC mixtures incorporating PHCM and that incorporating CA [a) MAC1, and b) MAC2] and NCA [c) MANC1, and d) MANC2] at different curing temperatures-------------------------------------------------------------------
363
Figure B-5 Relation between degree of hydration (amount of BW) and compressive strength development for the tested mixtures-------------
364
Figure B-6 Heat of hydration for UHPC mixtures incorporating a) PHCM, b) CA, and c) NCA--------------------------------------------------------------
365
Figure B-7 Drying shrinkage for a) PHCM mixtures and b) mixtures incorporating accelerating admixtures-------------------------------------
367
Figure B-8 Mass loss for a) PHCM mixtures, and b) mixtures incorporating accelerating admixtures------------------------------------------------------
367
Figure B-9 Autogenous shrinkage for a) PHCM mixtures and b) mixtures incorporating accelerating admixtures-------------------------------------
368
Figure D-1: Heat of hydration for mixtures incorporating wollastonite microfibers with a) 8% and b) 12% contents compared to that of the control mixture----------------------------------------------------------------
374
Figure D- 2: Total shrinkage for mixtures incorporating wollastonite microfibers with a) 8% and b) 12% contents compared to that of control mixture
375
Figure D- 3: Autogenous shrinkage for mixtures incorporating wollastonite microfibers with a) 8% and 6) 12% contents compared to that of the control mixture----------------------------------------------------------------
376
xxvii
LIST OF APPENDICES
Appendix A --------------------------------------------------------------------------------- 354
Appendix B --------------------------------------------------------------------------------- 355
Appendix C --------------------------------------------------------------------------------- 373
Appendix D --------------------------------------------------------------------------------- 374
Appendix E --------------------------------------------------------------------------------- 377
xxviii
Chapter 1 1
CHAPTER ONE
INTRODUCTION
1.1. ULTRA HIGH PERFORMANCE CONCRETE (UHPC)
Over the last few decades, concrete technology has experienced substantial advances,
resulting in innovative uses and unconventional applications of concrete. The use of
supplementary cementitious materials and additives has developed new generations of
concrete with enhanced properties, which can be used in areas that were dominated by
metals and ceramics. These new generations of concrete can be categorized based on
compressive strength development. Starting by Normal Concrete (NC) (20 to 40 MPa)
passing by high-performance concrete (HPC) (40 to 80 MPa), very high performance
concrete (VHPC) (100 to <150 MPa) and ultra high performance (≥ 150 MPa), which
represents a leap development in concrete technology.
In the early 1990’s, researchers at Bouygues company were the first to develop
ultra high-performance concrete (UHPC) (Richard and Cheyrezy, 1995). UHPC is
characterized by a very specific mixture design, which gives it a superior performance
compared to that of conventional concrete. The main concept behind UHPC mixture is to
minimize the number of defects, such as voids and internal micro-cracks, and to achieve a
greater percentage of the ultimate load capacity of its components (Acker and Behloul,
2004). This can be reached by enhancing homogeneity and increasing the packing density
Chapter 1 2
through optimization of the granular mixture and elimination of coarse aggregates
(Holschemacher and Weiße, 2005), producing a very dense and strong structure of the
hydration products using very low water-to-cement ratio (w/c) (about 0.20 to 0.25)
(Schmidt and Fehling, 2005).
The main characteristics of UHPC include very high compressive strength, a
relatively high tensile strength and enhanced durability compared even with that of HPC.
These outstanding properties make it a promising material for different concrete
applications (Tang, 2004). Nowdays, UHPC is used for producing special pre-stressed
and precast concrete members (Yazici, 2007). Applications include the production of
nuclear waste storage facilities (Yazici et al., 2009), precast pre-stressed concrete
highway bridge girders (Garas et al., 2009), pedestrian footbridges (Shah and Weiss,
1998), inner wedges and outer barrel of nonmetallic anchorage systems (Reda et al.,
1999), rehabilitation and retrofitting of concrete structures (e.g. waterproofing layer in
bridge decks, protection layer on a crash barrier walls and strengthening of industrial
floors) (Brühwiler and Denarié, 2008). Although there are only a few applications for
these concretes due to its high production cost, some economical advantages do exist in
UHPC applications (Yazici et al., 2009). For instance, it is possible to reduce
maintenance costs relative to steel and conventional concrete bridge girders (Garas et al.,
2009). Moreover, due to the ultra-high mechanical performance, the thickness of UHPC
elements can be reduced, which results in materials and cost savings (Yazici et al., 2009)
and increased useful space in buildings.
Indeed, it has been possible since several years ago to produce in the laboratory
concrete with a compressive strength as high as 700 MPa. But to reproduce it in a job site
Chapter 1 3
would be questionable due to uncontrolled in-situ conditions. Therefore, understanding
the influence of in-situ environmental conditions on UHPC behaviour is essential before
taking UHPC technology from the laboratory to full-scale field construction,
rehabilitation and retrofitting of concrete structures.
1.2. EARLY-AGE SHRINKAGE
In the literature, there is no agreed upon definition of the early-age period for concrete; it
highly depends on the investigated property. The early-age can be the first few hours or
days after casting concrete that are characterized by two main processes: setting
(progressive loss of fluidity) and hardening (gaining strength) (Pane and Hansen, 2002).
Generally, shrinkage can be divided into two main types: early-age shrinkage, which
represents shrinkage during the first 24 hours after the first contact between cement and
water, and can extend up to 7 days (Holt, 2001, Wongtanakitcharoen and Naaman, 2007,
Khan, 1995), and long-term shrinkage, which extends beyond the early-age period (Holt,
2001). The total shrinkage of concrete must be considered as the summation of both.
However, there is no clear relationship between the magnitudes of these two types of
concrete shrinkage.
Concrete shrinkage is a complex phenomenon that is affected by many factors
including drying and internal self-desiccation deformations, curing procedure and
ambient environmental conditions. Concrete shrinkage is generally assumed to begin at
the time of loading or drying, however, it actually starts as early as the cement and water
come in contact during concrete mixing. Early-age shrinkage is typically ignored in
Chapter 1 4
design codes of concrete structures since its magnitude is assumed to be much less than
long-term shrinkage (William et al., 2008). However, this can be erroneous since the
early-age shrinkage can be equivalent to or even more than the long-term one. New
generations of concrete are characterized by very low w/c (approximately below 0.40),
thus a high portion of measured shrinkage (i.e. up to 80%) can be a result of self-
desiccation shrinkage. Furthermore, a significant portion of this self-desiccation
shrinkage takes place during the early-age period (Holt, 2001).
Many serviceability and durability problems of concrete can be attributed to the
early formation of cracks as a result of concrete shrinkage. Recent studies have shown
that this can also significantly affect the initial term response of structures (Bischoff and
Johnson 2007). (Brown et al., 2007) reported that early-age cracking resulting from
drying shrinkage affects more than 100,000 bridge decks in the US. Therefore, several
shrinkage mitigation strategies have been proposed. These include adding shrinkage
reducing admixtures (SRA) (Tazawaa and Miyazawaa, 1995), controlling the binder
particle size distribution (Bentz et al., 2001), modifying the chemical composition of the
binder (Tazawaa and Miyazawaa, 1995), adding of saturated lightweight aggregates
(LWA) (Bentur et al., 2001) and superabsorbent polymers (SAP) (Igarashi and
Watanabe, 2006).
Substantial research has focused on predicting the early-age shrinkage behaviour
and evaluating the efficiency of various shrinkage mitigation methods under laboratory
conditions. Conversely, very limited research has explored that early-age shrinkage
behaviour under field-like conditions, whether with or without applying shrinkage
mitigation methods. This probably led to a misevaluation of the actual early-age
Chapter 1 5
shrinkage behaviour of different concrete mixtures and the real performance of shrinkage
mitigation methods.
1.3. NEED FOR RESEARCH
Despite the current knowledge and specifications for early-age shrinkage and mitigation
techniques, numerous early-age shrinkage cracking cases for different concrete structures
have been reported, indicating the existence of high shrinkage stresses (Brown et al.,
2007). It seems that the difference between the well defined curing conditions inside the
lab and the actual field conditions had contributed to a clear discrepancy between
laboratory measured performance and the actual performance of concrete mixtures in-
situ. Moreover, several studies have been conducted to quantify the shrinkage behaviour
of UHPC either under heat curing or constant temperature and relative humidity.
Therefore, a proper understanding for the effect of field exposure conditions still needs to
be gained, and the actual mechanisms that govern UHPC shrinkage of structural elements
cast in-situ is still largely unexplored.
On the other hand, several shrinkage mitigation techniques have been proposed
and are being used in various applications. The mechanisms of these shrinkage mitigation
techniques are based on certain chemical and physical processes that take place within
the cementitious material microstructure. Any factors that affect these processes are
expected to affect the efficiency and mitigation mechanisms of the used techniques.
Hence, a better understanding of the role and efficiency of these shrinkage mitigation
Chapter 1 6
techniques in UHPC, which exhibits more complex chemical and physical processes
compared to that of conventional concrete, under different exposure conditions is needed.
UHPC has a relatively high carbon-footprint and environmental impact due to its
high cement content, which leads to high energy consumption and CO2 emissions
associated with cement production. Hence, finding new strategies to produce more
environmentally friendly concrete, which exhibits lower or at least similar cracking
tendency to traditional cement-based materials, is our obligation to future generations
(Naik and Moriconi, 2005).
1.4. OBJECTIVES AND SCOPE
To address the aforementioned research needs, the fundamental theme of this research is
improving the current level of knowledge on the early-age shrinkage behaviour of UHPC
and evaluating the efficiency of shrinkage mitigation methods under realistic field
conditions. The specific objectives of the research are multi-fold:
• Introducing fundamental knowledge on the relationship between different types of
early-age shrinkage of UHPC under a wide range of simulated field conditions.
• Evaluating the influence of exposure conditions on the efficiency and mechanisms of
different shrinkage mitigation techniques.
• Implementing new environmentally-friendly shrinkage mitigation techniques that can
utilize the superior performance of UHPC along with achieving economical and
environmental benefits compared to that of convention mitigation techniques.
Chapter 1 7
• Investigating the synergetic effects of different shrinkage mitigation techniques in
order to identify optimum combinations that achieve efficient performance under
realistic field exposure conditions.
• Building modeling systems that are capable of predicting the early-age shrinkage of a
wide range of UHPC mixtures under various curing conditions.
To achieve the above objectives, the scope of this research includes:
• A wide range of simulated field-like conditions: the key variables included the
temperature (10°C, 20°C and 40°C) and ambient humidity (40, 60, 80 and 100%). In
the case of drying/wetting cycles, switching between a drying condition and a
submerging condition was conducted. For submerged conditions, specimens were
stored under water.
• Active shrinkage mitigation techniques: these include techniques that directly deal
with shrinkage sources and try to minimize their effects, including chemical
admixtures (i.e. shrinkage reducing admixture), since it reduces the surface tension of
the pore solution; and chemical additives (i.e. superabsorbent polymer) that acts as an
internal source of water.
• Passive shrinkage mitigation techniques: these include techniques that have indirect
effect on the developed shrinkage, including internal passive restraint systems, and
the addition of natural microfibers (e.g. wollastonite microfibers).
• Combined shrinkage mitigation techniques: since each shrinkage mitigation technique
has its benefits and drawbacks, combining two or more of the techniques above may
Chapter 1 8
optimize the gained benefits and minimize the drawbacks along with achieving better
shrinkage mitigation behaviour.
• Modelling: a model based on an artificial neural network (ANN), an emerging
computational intelligence-based tool in concrete technology research, was developed
to predict the shrinkage of the different types of concrete mixtures under the effect of
environmental exposure. The model was applied to the case of variable simulated
curing humidity and temperature including cold, normal and arid conditions. Based
on the experimental database and developed ANN model, a parametric analysis was
conducted to investigate the influences of different mixture criteria on the shrinkage
behaviour.
1.5. STRUCTURE OF THESIS
This thesis has been prepared according to the guidelines of the Faculty of Graduate
Studies at the University of Western Ontario for an integrated-article format. It comprises
11 chapters, 9 of which display the progress in the experimental program starting by
enhancing the fundamental knowledge about shrinkage and its mitigation techniques,
moving to introducing and evaluating new shrinkage mitigation techniques, followed by
modelling. Substantial parts of these chapters have been either published, accepted, or
submitted for possible publication in peer-reviewed technical journals and national and
international conferences.
Chapter 2 contains a state-of-the-art review of the current knowledge on early-age
properties of concrete, their relationships, influences of various shrinkage mitigation
Chapter 1 9
techniques and different proposed models. Chapter 3 discusses the interaction
mechanisms between different types of UHPC shrinkage under the effect of various
exposure conditions. In chapters 4 and 5, the role of different shrinkage mitigation
techniques and their performance under simulated field-like conditions are discussed.
Chapter 6 introduces a new concept for inducing an internal passive restraint to mitigate
UHPC shrinkage at the micro-structural level. The basic phenomena and experimental
results that demonstrate the concept are presented in this chapter. The performance of the
new shrinkage mitigation technique and its efficiency compared to that of conventional
mitigation techniques are described in Chapter 7. The benefits of adding different sizes
and contents of natural wollastonite microfibers to UHPC as partial volume replacement
for cement are presented in Chapter 8. Chapter 9 evaluates the performance of UHPC
incorporating different combinations of shrinkage mitigation techniques discussed in
previous chapters. The development of an artificial neural network model and parametric
analysis are presented in Chapter 10. Finally, general and specific conclusions drawn
from the research study along with recommendations for future research are included in
Chapter 11.
1.6. ORIGINAL CONTRIBUTIONS
This research introduces a series of fundamental investigations related to the early-age
shrinkage and the role of different shrinkage mitigation techniques under field-like
conditions. It explores the influence of a wide range of field-like conditions and the
efficiency of available shrinkage mitigation strategies. Moreover, it proposes new
Chapter 1 10
strategies to produce concrete with lower shrinkage and risk of cracking, yet with a
smaller environmental impact. Specific original contributions of this dissertation include:
1. Developing a large and comprehensive database on the early-age shrinkage of UHPC
under a wide range of environmental conditions.
2. Identifying the interaction mechanisms between different types of UHPC shrinkage
under simulated field-like conditions. Specifically, it was revealed that: (i)
autogenous and drying strains in UHPC specimens are dependent phenomena; (ii)
autogenous strains under sealed conditions differ from that under drying conditions;
(iii) applying the superposition principle without considering the effect of drying
conditions will result in overestimating autogenous strains.
3. Evaluating the performance and efficiency of different shrinkage mitigation
techniques and their interaction with the surrounding environmental. Specifically, it
was found that: (i) adequate external curing is essential for shrinkage mitigation
methods to work properly; (ii) drying conditions jeopardize the effect of SAP as a
shrinkage mitigating method; (iii) the efficiency of SRA admixtures is highly affected
by the ambient relative humidity, (iv) under submerged conditions, initial early
swelling results in very low net shrinkage strains; (v) the washout behaviour of SRA
under submerged and drying/wetting cycles was documented for the first time; (vi)
the washout of SRA can dismiss its effectiveness in mitigating shrinkage strains; and
(vii) combining SRA and SAP has a strong potential for developing a new generation
of high-performance shrinkage mitigation admixture with dual effect.
Chapter 1 11
4. Developing for the first time an innovative shrinkage mitigation technique capable of
improving the mechanical properties along with reducing shrinkage strains and
solving the economical and environmental problems associated with waste concrete.
More specifically: (i) the concept of self-retraining concrete was pioneered based on
using partially hydrated cementitious material (PHCM); (ii) it was demonstrated for
the first time that PHCM has a high potential for reducing shrinkage through
providing internal passive restraining system; (iii) left-over and/or returned concrete
can be used as PHCM, thus enhancing concrete sustainability and preventing waste
and disposal; (iv) PHCM showed a hydration acceleration effect, which has a
potential to replace conventional chloride-based accelerating admixtures, particularly
in the pre-cast industry; (v) combining PHCM and SRA mitigated the drawbacks of
SRA including delays in setting time and significant reduction in early-age
compressive strength; and (vi) combining PHCM and SAP reduced porosity and mass
loss during drying in UHPC.
5. Implementing the use of wollastonite microfibers in UHPC as a replacement of
cement, which has economical and environmental benefits along with enhancing the
early-age properties. In particular, (i) wollastonite microfibers enhanced the early age
compressive strength of UHPC mixtures incorporating SRA; (ii) wollastonite
microfibers promote pore discontinuity, leading to lower mass loss, drying shrinkage
and reduced SRA washing out; (iii) wollastonite microfibers act as passive internal
restraint leading to lower shrinkage; (iv) at early-age, the low elastic modulus of the
cementitious matrix increases the shrinkage restraining effect of wollastonite
microfibers; and (v) wollastonite microfibers delay the coalescence of micro-cracks.
Chapter 1 12
6. Advancing the promising use of artificial neural networks for predicting and
estimating the shrinkage behaviour of different concrete mixtures that are used in
concrete structures serving in different environments.
Chapter 1 13
1.7. REFERENCES
Acker, P. and Behloul, M., (2004), “Ductal ® technology: A large spectrum of properties, a wide range of applications,” Proceedings of the International Symposium on UHPC, Kassel, Germany, pp. 11-23.
Bentur, A., Igarashi, S. and Kovler, K., (2001), “Prevention of autogenous shrinkage in
high strength concrete by internal curing using wet lightweight aggregates,” Cement and Concrete Research, Vol. 31, No. 11, pp. 1587-1591.
Bentz, D.P., Jensen, O.M., Hansen, K.K., Olesen, J.F., Stang, H., and Haecker, C.J.,
(2001), “Influence of cement particle-size distribution on early-age autogenous strains and stresses in cement-based materials,” Journal of American Ceramic Society, Vol. 84, No. 1, pp. 129-135.
Bischoff, P.H., Johnson, R.D., (2007), “Effect of shrinkage on short-term deflection of
reinforced concrete beams and slabs,” In: Proceeding of structural implications of shrinkage and creep of concrete , Gardner, J., Chiorino, M.A. (eds.) ACI International, SP-246, pp. 167–180.
Brown, M.D., Smith, C.A., Sellers, J.G., Folliard, K.F. and Breen, J.E., (2007), “Use of
alternative materials to reduce shrinkage cracking in bridge decks,” ACI Materials Journal, Vol. 104, No. 6, pp. 629-637.
Brühwiler, E. and Denarié, E., (2008), “Rehabilitation of concrete structures using Ultra-
High Performance Fiber-Reinforced Concrete,” In: Proceeding of 2nd International Symposium on UHPC, Kassel, Germany, pp. 1-8.
Garas, V.Y., Kahn, L.F. and Kurtis, K.E., (2009), “Short-term tensile creep and shrinkage
of ultra-high performance concrete,” Cement and Concrete Composites, Vol. 31, No. 3, pp. 147-152.
Holschemacher, K. and Weiße, D., (2005), “Economic mix design of ultra high strength
concrete,” In: Proceedings of the 7th International Symposium on Utilization of High-Strength/ High-Performance Concrete, Washington D.C., Vol. 2, pp.1133-1144.
Holt, E.E., (2001), “Early age autogenous shrinkage of concrete,” Espoo 2001. Technical
Research Centre of Finland, VTT Publications 446, 184 p. Igarashi, S. and Watanabe, A., (2006), “Experimental study on prevention of autogenous
deformation by internal curing using super-absorbent polymer particles,” In: Proceedings of the International RILEM Conference - Volume Changes of Hardening Concrete: Testing and Mitigation, O.M. Jensen, P. Lura, and K. Kovler (eds.), RILEM Publications, pp. 77-86.
Chapter 1 14
Khan, A.A., (1995), “Concrete properties and thermal stress analysis of members at early-ages,” PhD Thesis, McGrill University, Canada.
Naik, T., R., Moriconi, G., (2005), “Environmental-friendly durable concrete made with
recycled material for sustainable concrete construction,” In Proceeding of international symposium on sustainable development of cement, concrete and concrete structures, Toronto, Canada, pp. 277-291.
Pane, I. and Hansen, W., (2002), “Early-age creep and stress relaxation of concrete
containing blended cements,” Materials and Structures, Vol. 35, No. 2, 92-96. Reda, M.M., Shrive, N.G. and Gillott, J.E., (1999), “Microstructural investigation of
innovative UHPC,” Cement and Concrete Research, Vol. 29, No. 3, 1999, pp. 323-329.
Richard, P. and Cheyrezy, M., (1995), “Composition of reactive powder concrete
Research,” Cement and Concrete Research, Vol. 25, No. 7, pp. 1501-1511. Schmidt, M. and Fehling, E., (2005), “Ultra-high-performance concrete: research,
development and application in Europe” In: Proceedings of the 7th International Symposium on Utilization of High-Strength/ High-Performance Concrete, Washington D.C., Vol. 1, pp. 51-77.
Shah, S.P. and Weiss, W.J., (1998), “Ultra high strength concrete; looking toward the
future,” In: ACI Special Proceedings from the Zia Symposium, Huston, Texas, USA.
Tang, M.C., (2004), “High performance concrete – past, present and future," In: Proceedings of the International Symposium on UHPC, Kassel, Germany, pp. 3-9.
Tazawaa, E. and Miyazawaa, S., (1995), “Influence of cement and admixture on
autogenous shrinkage of cement paste,” Cement and Concrete Research, Vol. 25, No. 2, pp. 281-287.
William, G.W., Shoukry, S.N. and Riad, M.Y., (2008), “Development of early age
shrinkage stresses in reinforced concrete bridge decks,” Mechanics of Time-Dependent Materials, Vol. 12, No. 4, pp. 343-356.
Wongtanakitcharoen, T. and Naaman, A.E., (2007), “Unrestrained early-age shrinkage of
concrete with polypropylene, PVA, and carbon fibers,” Materials and Structures, Vol. 40, No. 3, pp. 289-300.
Yazici, H., (2007), “The effect of curing conditions on compressive strength of ultra high
strength concrete with high volume mineral admixtures,” Building and Environment, Vol. 42, No. 5, pp. 2083-2089.
Chapter 1 15
Yazici, H., Yardımcı, H.Y., Aydın, S. and Karabulut, A., (2009), “Mechanical properties of reactive powder concrete containing mineral admixtures under different curing regimes,” Construction and Building Materials, Vol. 23, No. 3, pp. 1223-1231.
Chapter 2 16
CHAPTER TWO
EARLY-AGE PROPERTIES OF CONCRETE: OVERVIEW*
The long-term performance of concrete structures is affected to a large extent by the
properties and behaviour of concrete at early-age. However, the fundamental mechanisms
affecting the early-age behaviour of concrete have not yet been fully understood. This is
due to the various highly interrelated factors influencing it, and the complexity of testing
techniques needed for its investigation. With modern developments in concrete
technology, it has become essential to evaluate the influence of these interrelated factors
and their implications for the service life of concrete structures. Thus, a more
fundamental approach for investigating the early-age behaviour of concrete, along with
more reliable testing techniques is required. This chapter provides a critical overview of
research on the mechanisms that affect the properties of concrete and its performance at
early-age. It provides useful, concise and coherent information on the behaviour of
concrete at early-age, which should enhance the understanding of the implications of
such behaviour on the service life performance of concrete structures.
2.1. INTRODUCTION
Concrete is the most widely used construction material in the world. Its service life is
considered synonymous to its mechanical strength, durability and serviceability. The
selection of proper ingredients and mixture proportions are important to produce concrete
*A version of this chapter has been published in the Journal of Construction Materials, ICE, Vol. 164, Issue CM2, April 2011, pp. 55-77.
Chapter 2 17
that can meet strength and durability requirements. However, achieving a high quality
concrete may be elusive if adequate attention is not paid to its early-age properties. For
instance, the mechanical performance and long-term durability of concrete can be
seriously compromised due inadequate curing or insufficient compaction during
placement. Moreover, accelerated construction schedules aiming at economic gains have
led to tragic failures during construction due to inadequate knowledge of the concrete
behaviour at early-age (Oluokun et al., 1990). Therefore, a fundamental understanding of
the behaviour of concrete at early-age is essential to assure safety during construction, as
well as adequate durability and long-term properties.
There is no agreed upon definition of the early-age period for concrete; indeed it
depends on the investigated property. In other words, the time required to achieve a
certain level of a desired property is perceived as the early-age (Mehta and Monteiro,
2006). Table 2-1 summarizes the early-age period as considered by different researchers
(Mehta and Monteiro, 2006, RILEM, 1981, Kahouadji et al., 1997, Holt, 2001,
Wongtanakitcharoen and Naaman, 2007, Brooks and Megat-Johari, 2001, Holt and
Leivo, 2004, Altoubat, 2002, Østergaard et al., 2001, Khan, 1995, Oluokun et al.,1991,
Nassif et al., 2003, Kovler et al., 1999, Pane and Hansen, 2002, Bissonnette and Pigeon,
1995). Generally, the early-age is the first few hours or days after casting concrete that
are characterized by two main processes: setting (progressive loss of fluidity) and
hardening (gaining strength). During these processes, the fluid multiphase structure of the
fresh concrete transforms into a hardened structure due to the progress of hydration
reactions, leading to the development of mechanical properties, heat liberation and
deformations (Pane and Hansen, 2002). This heat liberation and water loss, due to either
Chapter 2 18
evaporation or consumption by hydration reactions, leads to internal/external
deformations. Hence, the coupling between thermal and mechanical characteristics of
early-age concrete is more critical compared to that in mature concrete. Furthermore,
proper curing after placement is crucial to maintain a satisfactory moisture content and
adequate temperature in concrete during this early stage so that the desired properties can
develop later (Huo and Wong, 2006).
Table 2-1: Definition of early-age period.
Early‐age period Field of study Reference. First 1‐2 days Concrete properties (Mehta and Monteiro, 2006) Up to 48 hrs Concrete properties (RILEM, 1981)
4 days after casting Strength (Kahouadji et al., 1997)
Shrinkage (Holt, 2001, Wongtanakitcharoen
and Naaman, 2007) Shrinkage and creep (Brooks and Megat‐Johari, 2001)
24 hrs after placing concrete
Creep and relaxation (Holt and Leivo, 2004) Shrinkage and creep (Altoubat, 2002)
5 days after casting Creep (Østergaard et al., 2001)
3 days after placing concrete
Strength (Khan, 1995, Oluokun et al.,1991)
Shrinkage (Nassif et al., 2003) Strength (Kovler et al., 1999)
Creep and relaxation (Pane and Hansen, 2002) Up to 7 days after
casting Creep (Bissonnette and Pigeon, 1995)
2.2. EARLY-AGE PROPERTIES
2.2.1. Early-Age Thermal Properties
Because of the significant role of the heat of hydration at early-age, basic thermal data for
concrete are of a vital importance. These data are essential to predict the temperature and
thermally induced stress and strain distributions within a concrete member in order to
Chapter 2 19
avoid thermal cracking. Generally, the thermal properties of concrete reflect those of its
constituents, i.e. the cement paste (including cement, water, chemical and mineral
admixtures) and aggregates, and these depend on the mixture proportions and type of
constituents (RILEM, 1981).
2.2.1.1 Heat of Hydration
The amount and kinetics of heat generated by cement hydration is an important parameter
for predicting the temperature development and its distribution within a concrete
member. The hydration of Portland cement is a highly exothermic chemical reaction
(Neville, 1996). Usually, about one-half of the total heat of hydration is evolved between
1 and 3 days after mixing cement with water (Oluokun et al., 1990). At early-age, the rate
and total heat of hydration are mainly influenced by the type, total content, and chemical
composition of cement, the ambient temperature and the admixtures used (Khan, 1995).
Cements with a high Dicalcium Silicate (C2S) and/or Tetracalcium Aluminoferrite
(C4AF) content can usually be considered as low heat cements, while cements high in
Tricalcium Silicate (C3S) and Tricalcium Aluminate (C3A) typically exhibit high heat
liberation (RILEM, 1981). Table 2-2 shows the typical heat of hydration for each of the
main cement phases (McCullough and Rasmussen, 1999). For instance, the ASTM C150
(Standard Specification for Portland Cement ) limited the C3S and C3A contents for low-
heat cement (Type IV) to 35% and 7%, respectively (ASTM, 2007). Furthermore, the use
of mineral admixtures, such as blast furnace slag (BFS), fly ash (FA) and other
supplementary cementing materials (SCM), in combination with ordinary portland
cement (Type I) in a blended cement, have proven to be a cost-effective method for
controlling the heat of hydration, which led to stopping the manufacture of the low heat
Chapter 2 20
cement Type IV. The hydration process in blended cements is relatively more complex
than that of the ordinary cement. It involves hydration reactions of the mineral additives
(pozzolanic reactions) resulting in heat liberation, in addition to that of the portland
cement hydration (Pane and Hansen, 2005). Generally, the total amount of heat liberated
will depend on the pozzolanic activity and proportion of the added mineral (Snelson et
al., 2008). For example, adding silica fume can accelerate the hydration of cement
resulting in a higher rate of heat of hydration, while adding BFS usually exhibit an
opposite trend (Alshamsi, 1997).
Table 2-2: Heat of hydration of cement compounds [after (McCullough and
Rasmussen, 1999), (Lea, 2004)].
Compound Heat of Hydration (J/g)
C3S (Tricalcium Silicate) 500‐520 C2S (Dicalcium Silicate) 260 C3A (Tricalcium Aluminate) 850‐910 C4AF (Tetracalcium Aluminoferrite) 420 C (Free Lime) 1165 MgO (Magnesium Oxide) 850
The ambient temperature has a significant effect on the rate of hydration and heat
liberation during early-ages. Figure 2-1 illustrates the effect of the temperature on the
heat of hydration for different types of Portland cement. In hot weather, the hydration rate
increases leading to more rapid heat generation. Conversely, in a cold climate, the
hydration rate slows down and lower heat is liberated. Hence, rapid hydrating cements
are used in such-conditions to avoid excessive set retardation and strength gain delays.
Moreover, chemical admixtures that affect the kinetics of hydration reactions normally
change the rate of heat liberation (Schindler, 2004). For instance, retarding agents slow
Chapter 2 21
the rate of hydration reactions at early-age; consequently, the rate and amount of heat
liberated will decrease. Conversely, accelerating admixtures increase the hydration rate.
However, the total long-term heat generated usually remains almost unchanged (RILEM,
1981).
0
100
200
300
400
500
600
700
800
900
1000
4 ˚C (40 ˚F) 24 ˚C (75 ˚F) 32 ˚C (90 ˚F) 41 ˚C (105 ˚F)
Temperature
Hea
t of h
ydra
tion
(J/g
)
Type IType IIType IV
Figure 2-1: Heat of hydration developed after 72 hours at different temperatures
(Type I: Ordinary Portland cement, Type II: Moderate sulfate resistance portland
cement, Type IV: Low heat portland cement) [adapted after (Neville, 1996)].
There are generally three commonly used methods for evaluating the heat
liberated from the hydration of cement (Khan, 1995), namely: conduction calorimetry,
heat of solution, and the adiabatic/semi-adiabatic method. Conduction calorimetry
measures the rate of heat released under nearly isothermal conditions. This method is
preferable for determining the heat of hydration at very early-ages (even less than 0.5
hour). For the heat of solution method, the heat of hydration is determined by the
Chapter 2 22
difference between the heat of solution of the dry cement and a hydrated cement paste
immersed in acid, respectively. This method is not suited for early-ages because of the
limited amount of heat being liberated. The adiabatic/semi-adiabatic methods measure
the temperature rise under insulated conditions, which simulates curing conditions at the
center of large concrete elements at early-ages. Generally, in the adiabatic method the
temperature of the sample surrounding environment must be equal to the temperature of
the concrete at any time. This condition requires that additional heat be supplied from the
outside. In the semi-adiabatic method, a maximum heat loss from the hydration specimen
less than 100 J/(h.K) is accepted. Therefore, the sample is insulated to minimize the rate
of heat loss (RILEM, 1997). The rate of hydration can be obtained from the product of
the specific heat, concrete density and the rate of change in the measured adiabatic
temperature curve.
2.2.1.2 Specific Heat Capacity
The specific heat (Cp) of a material is the amount of heat required to change a unit mass
of the material by one degree in temperature (Goldsmid, 1965). For concrete, Cp depends
on the mixture composition, moisture content and ambient temperature. The specific heat
of water is about 5 times greater than that of the un-hydrated cement and most aggregates
(Khan, 1995). Thus, the moisture content of concrete has a major effect on its specific
heat capacity at early-age as it changes significantly due to hydration and drying. A linear
reduction of the specific heat with the degree of hydration was reported (Bentur, 2003).
Generally, a decrease in the specific heat of concrete with age was consistently reported
by several researchers, while the opposite trend was not reported (Bentur, 2003).
However, there has been no general agreement concerning the magnitude of this decrease
Chapter 2 23
due to differences in the testing procedures and materials used. Table 2-3 summarizes
values of decrease of specific heat capacity with time reported by different researchers
(RILEM, 1981, Brown and Javaid, 1970, Mounanga, 2004). Moreover, the specific heat
of concrete was found to increase by about 3 % to 3.5 % per 10°C increase in temperature
(RILEM, 1981). However, little attention has been paid to the effect of the curing
conditions, including the ambient temperature and relative humidity on Cp. The curing
condition is expected to have a significant effect on the specific heat of concrete as it
directly affects the concrete moisture content and temperature. Therefore, a more
comprehensive study of the effects of different curing conditions on the development of
the specific heat of concrete at early-age is needed.
Table 2-3: Values of specific heat investigated by different researchers.
Age Initial Value
Age Later Value
Units Reduction
(%)* Sample Reference
3 days 1180 10 days
1160 ‐1.70 Mortar (RILEM, 1981)
24 hrs 1122 ‐2.52 6 hrs 1151
7 days 888 ‐22.85 Concrete
24 hrs 1055 ‐1.95 6 hrs 1076
7 days 871
J/kg per °C
‐19.05 Mortar
(Brown and Javaid, 1970)
6 hrs 2.6 24 hrs 2.35 106 J/m3 per K
‐ 9.62 Paste (Mounanga,
2004)
*Related to the Initial measured value.
2.2.1.3 Thermal Conductivity
The thermal conductivity (κ) is the rate of heat transmission across a unit cross-section
area when there is a unit temperature gradient perpendicular to that area, and is usually
Chapter 2 24
expressed as the ratio of the flux of heat to the temperature gradient (Goldsmid, 1965). It
significantly influences temperature gradients, thermal strains, warping, and cracking in
early-age concrete (Neville, 1996). A decrease in the thermal conductivity of concrete by
about 10 to 40% with age within the first week of hydration was observed (Østergaard,
2003). Conversely, other researchers stated that the thermal conductivity of concrete does
not change significantly with age, and may be considered constant (RILEM, 1981, Khan,
1995). The density, water content, temperature and mineralogical characteristics of
aggregates (see Table 2-4) significantly influence the thermal conductivity of concrete
(Khan, 1995). The thermal conductivity of cement paste is a resultant of the thermal
conductivity of its constituents, including un-hydrated cement, hydration products, and
that of the air and moisture contained in pores (Uysal et al., 2004). Air has a low thermal
conductivity; while water has about 25 times the thermal conductivity of air and less than
half that of the hydrated cement (Neville, 1996). Therefore, it is expected that the thermal
conductivity changes during the early period of hydration as the concrete phases change
(Khan, 1995). During this process, the air in pores is being partially displaced by water or
moisture; which in turn reacts with cement and is replaced by hydrated cement, leading to
greater thermal conductivity (Uysal et al., 2004). Therefore, variation in both the
moisture content and density of concrete can drastically affect the thermal conductivity of
concrete (Brown and Javaid, 1970). On the other hand, silica fume, latex, and
methylcellulose were found to decrease the thermal conductivity of concrete (Xu and
Chung, 1999, Fu and Chung, 1997). This is mainly due to the relatively low thermal
conductivity of these admixtures.
Chapter 2 25
Table 2-4: Thermal properties of concrete with different aggregate types [after
(Mehta and Monteiro, 2006)].
Aggregate Type Thermal
conductivity(W/m.K)
Thermal diffusivity(m2/h)
Coefficient of thermal expansion
(microstrain /°C) Quartzite 3.5 0.0054 ≈ 12.0 Dolomite 3.2 0.0047 ‐‐‐‐ Limestone 2.6‐3.3 0.0046 ≈ 6.0 Granite 2.6‐2.7 0.0040 ≈ 8.0 Rhyolite 2.2 0.0033 ‐‐‐‐ Basalt 1.9‐2.2 0.0030 ≈ 7.0
There are mainly three common methods for measuring the thermal conductivity
of concrete (Kook-Han et al., 2003), namely the two-linear parallel-probe (TLPP)
method, the plane-heat-source (PHS) method, and the hot-guarded plate (HGP) method.
The required sample preparation was the main obstacle for applying these methods to
concrete at early stages. However, a new experimental device based on the basic
principle of the TLPP method was recently developed. It showed reasonable reliability
for measuring the thermal conductivity of concrete at very early-age (Kook-Han et al.,
2003). Moreover, Bentz (Bentz, 2007) demonstrate the applicability of the transient plane
source method (TPS) for measuring the thermal conductivity of hardening concrete.
These developments should provide useful tools for investigating the thermal
conductivity of different concrete mixtures during early-ages and significant research
work is still needed in this area.
2.2.1.4 Thermal Diffusivity
The thermal diffusivity (kd) of a material represents the rate at which temperature changes
within its mass (Goldsmid, 1965). It measures the rapidity of heat propagation through a
Chapter 2 26
material. It is an important property in problems involving non-steady state heat
conduction. The thermal diffusivity of ordinary concrete generally depends on the
aggregate type used (see Table 2-4) and water content (Mehta and Monteiro, 2006,
Neville, 1996). Only limited and conflicting data are available on the thermal diffusivity
of concrete at early-age. Constant thermal diffusivity values through the hardening
process were observed, as well as increasing and decreasing trends (De Schutter and
Taerwe, 1995). Generally, the thermal diffusivity is related to the thermal conductivity
and specific heat by the following mathematical formula (Goldsmid, 1965):
pd C
k⋅
=ρκ
(m2/s) Eq. 2-1
Where, k is the thermal diffusivity (m2d /s), κ is the thermal conductivity (W/(m.K)), ρ is
the bulk density (kg/m3), and Cp is the specific heat (J/(kg.K)). Table 2-5 summarizes
values for the thermal diffusivity of concrete reported by various researchers (RILEM,
1981, Bentur, 2003, Brown and Javaid, 1970).
The conflicting data of thermal diffusivity of concrete can be attributed to
differences in the testing methods, mixture proportions and constituents of tested
concrete, and curing conditions. These factors have a significant effect on the degree and
rate of concrete phase development, which can lead to diverse thermal characteristics.
Even applying the above mathematical relationship (Eq. 2-1) is expected to give different
thermal diffusivity values since it depends on the thermal conductivity and specific heat,
which do not have a well-defined trend during early-age.
Chapter 2 27
Table 2-5: Values of thermal diffusivity obtained by different researchers.
Age Value (m2/hr) Trend Sample Reference
1 day 0.00370 7 days 0.00370
Constant Concrete (RILEM, 1981)
1 hr 0.00336 7 days 0.00283
Decrease (‐15.85%) Concrete
1 day 0.00255 7 days 0.00296
Increase (16.08%) Mortar
(Brown and Javaid, 1970)
0.85000 Within 1000 hrs 1.05000
Increase (23.53 %) Paste (Mounanga,
2004)
2.2.1.5 Coefficient of Thermal Expansion
The coefficient of thermal expansion (CTE) of a material is the change in a unit length of
the material per one degree of temperature change (Goldsmid, 1965). The CTE of
concrete depends on different parameters, including the type and content of cement,
aggregate type (see Table 2-4), w/c, age, temperature, and relative humidity (Østergaard,
2003). The CTE of concrete can be estimated from the volumetrically weighted average
of the coefficients of thermal expansion of its mixture components (Kada et al., 2002).
At early-age, especially during the first ten hours, an important variation in the
CTE of concrete occurs, and then it stabilizes (see Fig. 2-2). This variation is essentially
caused by variation in the amount of water that is not yet chemically bound since water
has a CTE up to 20 times greater than that of other concrete constituents (Kada et al.,
2002). Hence, it is believed that the CTE of concrete at early-age, when concrete
typically has a higher free water content, is several times higher than that of the hardened
concrete (Østergaard, 2003, Kada et al., 2002), and can be estimated to be 3 to 4 times
that of the hardened concrete at the onset of setting, then decreases until the end of the
Chapter 2 28
setting period to remain nearly constant thereafter (Kada et al., 2002). Table 2-6
summarizes CTE values of concrete as a function of its age (RILEM, 1981, Khan, 1995,
Bentur, 2003, Østergaard, 2003, Kada et al., 2002, Cusson and Hoogeveen, 2006,
Sellevold and BjØntegaard, 2006). Generally, a decreasing trend of CTE with age was
observed, however its amount is highly varied. This can be attributed to the ongoing
hydration, change in primary constituents and surrounding conditions, which affect the
amount of free water and proportion of concrete constituents at the testing age.
0
5
10
15
20
25
30
35
40
5 10 15 20 25 30 35 4Age (hours)
CTE
(mm
/°C)
0
w/c=0.30
w/c=0.35
w/c =0.45
Figure 2-2: Evaluation of the coefficient of thermal expansion of concrete during
hydration [adapted after (Kada et al., 2002)].
Chapter 2 29
Table 2-6: Values of coefficient of thermal expansion obtained by different
researchers.
Age / °C × 10‐6 Reference 4 hrs 70.00
≥120 hrs 12.00 (Østergaard, 2003)
7 hrs 24.00 ≥ 13 hrs 10.00
(Khan, 1995)
1 day 6.00 4 days 8.00
(Cusson and Hoogeveen, 2006)
Fresh concrete 20.00 8‐24 hrs 15.00 1 ‐6 days 12.00
(RILEM, 1981)
10‐24 hrs ≈ 9.10 to 12.30 1 day – 4 days ≈ 8.50to 8.10
> 4 days ≈ 8.00 (Bentur, 2003)
8 hrs ≈ 22.11 12 hrs ≈ 7.12 ≥ 1 day ≈ 5.79
(Kada et al., 2002)
Before setting 20.00 Final set 7.0011 weeks 11.00> 11 weeks 7.00
(Sellevold and BjØntegaard, 2006)
In addition, during the hydration period, the accuracy of measuring small
displacements in concrete is strongly dependent on its stiffness, which presents the main
difficulty during the testing stage. On the other hand, dealing with a heterogeneous,
porous and aging material such as concrete, presents other difficulties during the analysis
stage. Thus, a special extensometer with a sufficiently low stiffness (1 GPa) was used to
monitor the concrete strain at very early-age (Kada et al., 2002). This method allows
monitoring the evolution of the CTE from the onset of concrete hardening. In order to
maximize the accuracy of readings, an experimental approach to determine the
coefficient of thermal expansion of concrete at early-age under stress-free and isothermal
Chapter 2 30
conditions using high accuracy LVDTs was developed (Cusson and Hoogeveen, 2006).
As a result, the evolution of CTE as a function of time after casting could be monitored.
Nevertheless, it was not possible to accurately measure the CTE before the final setting.
(Viviani et al., 2006) introduced a system of optical fiber sensors to monitor the evolution
of CTE based on determine the kinetic parameters of hardening materials. However the
measurement system allows rapid and efficient predictions of CTE under laboratory
conditions, its potential for in-situ application is still need to be investigated (Viviani et
al., 2006).
2.2.2. Early-Age Mechanical Properties
The mechanical properties of concrete at early-age depend, to a large extent, on the
development of the hydrated cementitious microstructure as a function of the achieved
degree of hydration. However, the rate of development differs from one mechanical
property to another. Since the interrelations between these mechanical properties are
affected by numerous factors such as the mixture proportions, w/c, age, curing
conditions, rate of loading, it is difficult to formulate a unique relationship between such
properties.
2.2.2.1 Compressive Strength
The compressive strength is considered as a key property of concrete. It provides a
general indication of concrete quality. The gain in compressive strength is typically rapid
at early-age, and then becomes slower at later ages. This early rapid increase in strength
is directly related to the increase in the calcium silicate hydrate (CSH) gel/space ratio
(Neville, 1996). Compressive strength is influenced by several factors; most notably the
Chapter 2 31
w/c, type of cement, additives, and curing conditions (Østergaard, 2003). Rapid hydration
of cement results in a higher degree of hydration and consequently higher early-age
strength for a given w/c ratio (Mehta and Monteiro, 2006).
Curing conditions, i.e. the availability of moisture and the temperature profile,
drastically affect the compressive strength gain. At a very early-age, the absence of
moisture usually has a limited effect on the early strength gain, because concrete is still
wet. However, inadequate moisture curing during the first day after casting concrete
could lead to noticeable strength loss at later ages (RILEM, 1981). Furthermore, a high
initial curing temperature speeds up the hydration reactions and the formation of the
hydrated cement paste structure at early-age. Thus, it enhances the early-age compressive
strength of concrete. However, it decreases the strength at later ages due to the lower
quality of hydration products microstructure formed at higher temperature (Kahouadji et
al., 1997).
Pozzolanic materials can also contribute to the early-age strength through
improving the particle packing density (filler effect) and densifying the aggregate-cement
paste interfacial transition zone (Neville, 1996). In addition, the type and level of addition
of pozzlans have a significant influence on the compressive strength development. For
instance, concrete incorporating class C (high calcium oxide) fly ash generally develops
higher early-age strength than that of concrete made with class F (low calcium oxide) fly
ash (Kosmatka et al., 2002). Moreover, other pozzolanic materials such as blast furnace
slag were found to slightly improve the early-age strength compared to its more
significant contribution to the later strength, which can be ascribed to its slow hydration
rate (Shan-bin and Zhao-jia, 2002). However, applying the crystal seed technology, i.e.
Chapter 2 32
addition of hydrated micro-crystals, for this type of pozzolanic materials was found to
improve its early strength contribution significantly (Prusinski, 2006). Generally, using
pozzolanic materials may result in higher or lower early compressive strength of concrete
depending on their type, addition rate, mineralogy, particle shape and fineness, and
pozzolanic activity.
Several methods have been developed to predict the compressive strength of
concrete at early-age. One of the well-known methods is the maturity or the “equivalent
age” method, which is expressed as a function of the time and temperature of curing as
follows:
∑ ∆⋅−= tTTtM a )()( 0 Eq. 2-2
Where ∆t is the time interval, Ta is the average concrete temperature during the time
interval ∆t, and T0 is the datum temperature (Mehta and Monteiro, 2006).
At early-ages, maturity is low, especially in the first two days after casting, and
the relation between compressive strength and maturity is not fully linear. Thus, for
accurate compressive strength prediction at early-age using the maturity concept, the heat
of hydration should be considered during the calculation of maturity (Oluokun et al.,
1990). Figure 2-3 shows the average additional maturity due to the heat of hydration
with respect to concrete strength (σc). Moreover, a more fundamental approach to
evaluating the compressive strength development as a function of the degree of
hydration, taking into account the w/c as a parameter, was introduced (De Schutter and
Taerwe, 1996). Good agreement with the available data was obtained based on following
formula:
Chapter 2 33
( )( )
a
o
o
c
c
rrr
rfrf
⎟⎟⎠
⎞⎜⎜⎝
⎛−−
== 11
Eq. 2-3
Where is the compressive strength at degree of hydration r, is the
compressive strength at a degree of hydration r=1, r
( )rfc ( 1=rfc )
o and a are parameters that depend on
the cement type and w/c (De Schutter and Taerwe, 1996).
0
200
400
600
800
1000
1200
1400
6 hr 12 hr 1 day 2 days andabove
Age
Mat
urity
(F-h
r)
7.5-10.0 kips (52.0-69.0 MPa)5.0-7.5 kips (35.0-52.0 MPa) 2.5-5.0 kips (17.2-35.0 MPa)
σc=
σc=
σc=
Figure 2-3: Average additional maturity, (F-hr), due to hydration [adapted after
(Oluokun et al., 1990)].
2.2.2.2 Tensile Strength
Tensile strength is a key property of early-age concrete; it has a chief influence on the
resistance of concrete to plastic shrinkage, thermal stresses during hydration, and early-
age loading and cracking which can affect the stiffness of the structure (Zhao, 1990).
Measuring the early-age tensile strength of concrete is complicated since concrete is
Chapter 2 34
viscous and inelastic at that stage (Abel and Hover, 1998). Several methods have been
developed in an attempt to evaluate this property including uniaxial tensile test, splitting
test, and flexural test. Results obtained by the uniaxial tensile test can be characterized as
the real tensile strength of concrete. However, the specimen’s self-weight for the vertical
testing position or the friction against the sub-base for the horizontal testing position
induce a significant error (Østergaard, 2003). The splitting tensile test (Brazilian test) and
the three point flexural test rely on linear elastic formulas with empirical correction
factors, which make their applicability for early-age concrete questionable. Fracture
mechanics testing methods, e.g. the crack mouth opening displacement (CMOD), have
increasingly been applied to evaluate the early-age tensile strength (Østergaard, 2003).
The development of tensile and compressive strength is generally affected by
similar factors. Thus, the tensile strength of concrete can be related to its compressive
strength. This relationship is influenced by age, grading, type and density of aggregates,
curing conditions, and the strength evaluation method. At early-age, the tensile strength
tends to increase more rapidly than the compressive strength (RILEM, 1981, Bentur,
2003, De Schutter and Taerwe, 1996). Conversely, some researcher found that the tensile
strength increases at a lower rate than that of the compressive strength
(Swaddiwudhipong et al., 2003), and that the ratio between the two properties decreases
from 0.1 to 0.04 as the concrete matures. This discrepancy can be attributed to
differences in the starting age of the test and quality of concrete. (Kasai et al., 1972)
reported a higher increasing rate for the tensile strength compared to that of the
compressive strength during the very early age up to around 0.5 day. Thereafter, the
Chapter 2 35
tensile strength increasing rate became lower than that of the compressive strength, as
shown in Fig. 2-4.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0 0.5 1 1.5 2 2.5 3 3.5Age (Day)
Tens
ile/C
ompr
essi
ve S
tren
gth
Rat
io
w/c=0.6 (Swaddiwudhipong et al., 2003)
w/c=0.4 (Swaddiwudhipong et al., 2003)
w/c=0.45 (RILEM, 1981)
w/c=0.4 (Kasai et al., 1972)
w/c=0.6 (Kasai et al., 1972)
Figure 2-4: Tensile / compressive strength of concrete versus age.
The tensile strength can be expressed as a function of the compressive strength as
follows (Neville, 1996):
nct fkf )(⋅= Eq. 2-4
Where is the tensile strength, is the compressive strength, and k and n are relation
coefficients as shown in Table 2-7 (Zhao, 1990, ACI, 2005, Raphael, 1984, CEB-FIP,
1991). This relationship was obtained for matured concrete. Yet, previous research has
indicated its validity at early-ages and under various conditions (Zhao, 1990).
tf cf
Chapter 2 36
Table 2-7: Different suggested values for k and n.
Direct tensile strength Modulus of rupture Splitting tensile strengthk 0.5 0.7 0.3 0.95 6.7 0.33n 0.5 2/3 2/3 2/3 0.5 2/3
Ref. (Zhao, 1990)
(Raphael, 1984)
(CEB‐FIP, 1991)
(Raphael, 1984)
(ACI, 2005) (CEB‐FIP, 1991)
The tensile strength evolution between the ages of 2 and 8 hrs after mixing
concrete was investigated by (Abel and Hover, 1998). A dormant period of tensile
strength gain was observed from the age of 2 hrs to 4 hrs at which tensile strength was
infinitely low. This period was followed by a very rapid tensile strength development
starting at the point of initial setting. However, it is difficult to monitor this dormant
period for mixtures characterized by a delayed setting. On the other hand, the tensile
strain capacity (strain at maximum stress) starts with a very high value before the initial
setting and continues to decrease beyond that, reaching a minimum value later (Fig. 2-5)
(Hammer et al., 2007).
Chapter 2 37
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
8 9 10 11 12Age (hours)
Tens
ile S
tren
gth
(MPa
)
Strain rate :0.002 mm/s
Strain rate :0.16 mm/s
(a)
0
2000
4000
6000
8000
10000
8 9 10 11 12Age (hours)
Stra
in C
apac
ity (1
0 -6
)
Strain rate :0.002 mm/s
Strain rate :0.16 mm/s
(b)
Figure 2-5: a) Uniaxial tensile strength, and b) Uniaxial tensile strain capacity
[adapted after (Hammer et al., 2007)].
Chapter 2 38
Decreasing the w/c was reported to increase the early-age tensile strength
development (Fig. 2-6) (Abel and Hover, 1998). Using shrinkage reducing admixtures
(SRA) was found to have a non significant effect on the early-age tensile strength of
concrete at different w/c as shown in Fig. 2-7 (D’Ambrosia et al., 2001, D’Ambrosia et
al., 2002).
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 2 4 6 8 10Time after Batching (hours)
Tens
ile S
tren
gth
(MPa
)
0
20
40
60
80
100
120
Tens
ile S
tren
gth
(psi
)
w/c=0.3w/c=0.4w/c=0.5w/c=0.6w/c=0.7
Figure 2-6: Tensile strength of concrete with different water/cement ratio [after
(Abel and Hover, 1998)].
Chapter 2 39
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0 2 4 6 8 10Time (days)
Tens
ile S
tren
gth
(MPa
)
0
100
200
300
400
500
600
Tens
ile S
tren
gth
(psi
)
w/c=0.35w/c=0.45w/c=0.35+1%SRAw/c=0.45+1%SRA
Figure 2-7: Tensile strength of concrete with shrinkage reducing admixture
[adapted after (D’Ambrosia, 2002)].
2.2.2.3 Modulus of Elasticity
The modulus of elasticity is a principle property of concrete; it indicates the concrete’s
capability to deform elastically and thus is related to its serviceability. Generally, it is
obtained from the stress-strain curve at a certain stress level relative to the ultimate
strength (Tia et al., 2005). This stress value is 40% of the ultimate strength according to
ASTM C 469-94 (Standard Test Method for Static Modulus of Elasticity and Poisson's
Ratio of Concrete in Compression) (ASTM, 1994), and about 33% in the British
Standards (BS 1881: part 121:1983: Method for determination of static modulus of
elasticity in compression) (BSI, 1983). On the other hand, researchers have tried to
monitor the evolution of the elastic modulus at early-age using ultrasonic waves and to
formulate a relationship between the measured dynamic modulus values ( ) and the dE
Chapter 2 40
static modulus ( ), especially for high-performance concrete. A relationship between
the dynamic and elastic modules was given by (Mesbah et al., 2002):
cE
2.311 )160065(109 +×= −dc EE Eq. 2-5
The proposed correlation was obtained based on experiential results, which makes it
limited to the used aggregates and cement type. However, it indicates the possibility of
establishing a relationship between the concrete’s dynamic and static modules of
elasticity at early-age.
Several formulas have been proposed to express the modulus of elasticity of
concrete at any age as a function of its compressive strength (Table 2-8) (Zhao, 1990,
Ghali and Favre, 1994, Gardner and Zhao, 1993). However, the change in the concrete
elastic modulus with respect to its compressive strength at early-ages is affected by
several parameters, including the w/c, temperature; cement type, properties of aggregates,
and curing conditions (Neville, 1996). Decreasing the w/c enhances the modulus of
elasticity as a result of developing a denser microstructure (Østergaard, 2003). The age of
concrete at the time of testing was not found significantly to affect the relationship
between the modulus of elasticity and compressive strength as shown in Fig. 2-8 (Zhao,
1990). This is due to the very rapid development rate of the modulus of elasticity of
concrete at early-age compared to that of compressive strength. About 90% or more of
the elastic modulus value is achieved within the first 24 hours after casting (Myers,
1999). Consequently, the risk of cracking at early-age rises since the stress generated
depends on the modulus of elasticity, whereas the resistance to cracking depends on the
development of tensile strength (Atrushi, 2003).
Chapter 2 41
Table 2-8: Modulus of elasticity as a function of compressive strength of concrete at
any age.
Equation Reference
⎟⎠⎞
⎜⎝⎛
+′=
ttftE o
c 85.0457000)( 0
Where E (t0) = modulus of elasticity at age t0, = compressive strength
(psi), and t = age of concrete. cf ′
(Ghali and Favre, 1994)
cmc fE ′+= 43003500
Where Ec = time dependent modulus, = compressive strength (MPa) cmf ′
(Gardner and Zhao, 1993)
3/2)(0.3 cmc fE ′= for MPa 27≤′cmf3/1)(0.9 cmc fE ′= for MPa 27>′cmf
Where Ec = modulus of elasticity, cmf ′ = compressive strength at test
time.
( )28,,,28,, 2.112.008.08 cutcutcucutc ffffE ′′+′+′+=
Where Ec,t = modulus of elasticity at age t, = compressive strength
at 28 days, = compressive strength at age t days, and t ≥ 3 days.
28,cuf ′
tcuf ,′
(Zhao, 1990)
0
10
20
30
40
50
60
70
0 20 40 60 80Compressive Strength (MPa)
Mod
ulus
of E
last
icity
(GPa
)
Test age less than 3 daysTest age from 3 to 6 daysTest age from 7 to 14 daysTest age from 15 to 30 daystest age more than 30 daysPoly. (Series1)
Figure 2-8: Modulus of elasticity to compressive strength ratio data classified by
concrete age [after Zhao, 1990].
Chapter 2 42
2.2.2.4. Poisson’s Ratio
Poisson’s ratio is the ratio between the lateral and longitudinal strains under the same
uniaxial load (Neville, 1996). Limited research focused on the evolution of the Poisson’s
ratio of concrete at early-ages. A review of previous studies showed that the Poisson’s
ratio changes with time. It starts with a high value at early stage, then it decreases sharply
until it reaches a low value during the first 24 hours, then it starts to increase again
(Mesbah et al., 2002, Byfors, 1980). Conversely, other researchers stated that the
Poisson’s ratio of concrete does not change significantly with age, and may be considered
constant (Oluokun et al., 1991). Considering the assumption that the deformation of fresh
concrete occurs without volume changes, the Poisson’s ratio can be taken equal to 0.5
(De Schutter and Taerwe, 1996). Table 2-9 gives typical Poisson’s ratio values observed
by different researchers (RILEM, 1981, Mesbah et al., 2002, Byfors, 1980, Oluokun et
al., 1991). It was found that the Poisson's ratio does not change significantly with the
increase of the cement content (Oluokun et al., 1991, Oluokun, 1989). Moreover, tests
conducted on concrete specimens at different ages (1, 3, 7, 14, 28, 90 and 180 days) have
shown an insignificant effect of the moisture content and ambient temperature on
Poisson’s ratio (Downie, 2005). Generally, the Poisson’s ratio can be evaluated as a
function of the degree of hydration as follows:
( ) rerr ⋅−⋅+⎟⎠⎞
⎜⎝⎛ ⋅
⋅= 105.02
sin18.0 πν Eq. 2-6
Where ( )rν is the Poison’s ratio at a degree of hydration r (De Schutter and Taerwe,
1996).
Chapter 2 43
Table 2-9: Poisson’s ratio values.
Initial value Minimum value* Reference 0.23 0.14 (Mesbah et al., 2002) 0.40 0.20 (RILEM, 1981) 0.48 0.13 (Byfors, 1980)
0.19 (Oluokun et al., 1991) *Minimum measured value during the first 24 hr
2.2.3. Early-Age Shrinkage of Concrete
Shrinkage of concrete is the main cause for several types of cracks which influence the
serviceability and durability of concrete. Shrinkage can be divided into two main types:
early-age which represents shrinkage during the first 24 hours after the first contact
between cement and water (Holt, 2001, Wongtanakitcharoen and Naaman, 2007, Khan,
1995), and long-term shrinkage, which extends beyond the first day (Holt, 2001). The
total shrinkage of concrete must be considered as the summation of both. However, there
is no clear relationship between the magnitudes of these two types of shrinkage. At rapid
drying conditions, the magnitude of the early-age shrinkage can easily exceed that of the
long-term shrinkage. Figure 2-9 illustrates this behaviour for different environmental
conditions (Holt, 2005). Due to its low strength and strain capacity at early-age, concrete
is very sensitive to internal stresses. Thus, any induced tensile stress greater than the
tensile strength will cause either cracks comparable to those that occur at later ages, or
internal and microscopic cracks which can be reactivated at later ages, causing serious
problems (Holt and Leivo, 2004). Stresses that lead to early-age cracking are typically the
result of a combination of three types of deformation: thermal dilation, drying shrinkage,
and autogenous shrinkage.
Chapter 2 44
Figure 2-9: Accumulation of early-age and long-term shrinkage, with various curing
environments (Wind = 2 m/s (4.5 mph), dry = 40% RH, wet = 100% RH) [after
(Holt, 2005)].
2.2.3.1 Thermal Dilation
Thermal dilation is induced by inadequate heat dissipation during the cement hydration
and cooling of concrete. It can be observed as an expansion and/or contraction
deformation. During the hydration of cement at early-age, the liberated heat increases the
concrete temperature causing thermal expansion. As the hydration reactions reach its
peak, the rate of heat liberation decreases and concrete starts to dissipate heat and cools
down, causing contraction or thermal shrinkage (Holt, 2001). In massive concrete
however, the difference in the rate of heat dissipation between the interior and exterior
concrete can induce a thermal gradient, leading to thermal strains, and associated stresses
that can cause cracking (Holt, 2001, Neville, 1996). Because the coefficient of thermal
Chapter 2 45
expansion varies during early-ages, the traditional method of multiplying the temperature
change by an average value of the coefficient of thermal expansion is not accurate for the
evaluation of the early-age thermal dilation. Other approaches including the Poly-
isothermal test and Saw-toothed test were introduced (Bjøntegaard and Sellevold, 2001).
In these methods, the tested specimen is subjected to a realistic temperature history,
comprising of temperature steps. The basic idea behind these new approaches is that
thermal dilation can be measured directly during each temperature step. Therefore,
thermal dilation can be separated from other early-age deformations.
2.2.3.2 Drying Shrinkage
Drying shrinkage is the reduction in the concrete volume due to moisture loss at constant
temperature and relative humidity. It is mainly affected by the total moisture loss, rate of
evaporation and bleeding. If the bleeding rate is higher than the evaporation rate, drying
shrinkage will not occur since the excess water will be sufficient for further evaporation
and can act as a curing layer (Holt, 2001, Holt, 2005). When the reverse behaviour
occurs, the extra water required for evaporation will be extracted from the internal pores.
Losing water from capillary pores at early-age will cause plastic shrinkage and
subsequent internal stresses, leading to early-age surface cracks (Khan, 1995). The
particle size distribution and quantity of cement, w/c and aggregates have an essential
role in controlling bleeding, which in turn controls the mechanism of drying shrinkage on
the macro-level. However, it is believed that the drying shrinkage mechanism on the
nano-level is a combination of four well-known mechanisms (Bentur, 2003), namely:
surface free energy, capillary tension, movement of interlayer water, and disjoining
pressure. This combination was found to be highly affected by the relative humidity of
Chapter 2 46
concrete (Bentur, 2003), which varies drastically during early ages. In addition, it should
be noted that the measured shrinkage strain of drying concrete specimens under
isothermal conditions is not representative of the net drying shrinkage. This measured
shrinkage includes drying shrinkage and shrinkage due to cement hydration (so-called
autogenous shrinkage), especially at low w/c concrete mixtures.
2.2.3.3 Autogenous Shrinkage
Autogenous shrinkage of concrete can be defined as the macroscopic volume change that
occurs after the initial setting as a result of the withdrawal of moisture from capillary
pores to continue cement hydration reactions (Mihashi and Leite, 2004). Chemical
shrinkage, which is the reduction in the volume of hydration products compared with that
of the reacting constituents, can be considered as the main driving mechanism behind
autogenous shrinkage (Tazawa, 1999). However, autogenous shrinkage is considered as
an external volume change (apparent volume change), while chemical shrinkage is
considered as an internal volume reduction (absolute volume change), as shown in (Fig.
2-10) (Holt, 2001, Mihashi and Leite, 2004, Tazawa, 1999, Esping, 2007).
Chapter 2 47
Figure 2-10: Chemical shrinkage and autogenous shrinkage of concrete[adapted
after Mihashi and Leite (2004)].
During early-age, concrete undergoes three phases including particulate
suspension, skeleton formation, and initial hardening (Fig. 2-11). At the initial phase,
concrete is still plastic without a permanent internal structure, and autogenous shrinkage
is equivalent to chemical shrinkage (part AB in Fig. 2-11). Consequently, any applied
stresses will cause movement of the body. A few hours after casting, the development of
a skeleton starts due to the formation of hydration products. During this stage, the setting
will occur, and concrete can resist some of the chemical shrinkage. At this moment, the
so-called mineral percolation threshold, autogenous shrinkage starts to diverge from
chemical shrinkage (part BC in Fig. 2-11). Finally, the hardening stage starts and
autogenous shrinkage becomes increasingly restrained due to stiffening of the cement
paste (part C in Fig. 2-11) (Holt, 2001, Bentur, 2003, Tazawa, 1999, Esping, 2007).
Chapter 2 48
Figure 2-11: Autogenous and chemical shrinkage during different stages, as a
function of degree of hydration [adapted after (Holt, 2001, Esping, 2007)].
Measuring autogenous shrinkage should start as soon as possible after mixing the
concrete since there is no general agreement on the starting point of autogenous
deformation (the time zero (t0)) (Esping, 2007). While autogenous shrinkage has been
known for a long time, it’s practical importance and effects have only been recognized
with the development of low w/c concrete mixtures (w/c below 0.42) (Aitcin, 2003).
Autogenous shrinkage can have nearly the same value as drying shrinkage for low w/c
concrete under normal condition (20°C and RH=50%) (Tazawa and Miyazawa, 1995).
Several methods for measuring the early-age shrinkage of concrete have been
reported by several researchers. It is believed that the difference between the various
measuring methods is largely responsible for variations in results. Table 2-10
summarizes the common methods for measuring early-age shrinkage and their main
Chapter 2 49
features (Tazawa, 1999, Bjøntegaard et al., 2004, Lee et al., 2003, Termkhajornkit et al.,
2005, Hansen and Jensen, 1997, Lokhorst, 1999, Glišic and Simon, 2000, Jensen and
Hansen, 1995, Østergaard and Jensen, 2003, Gagné et al., 1999, Lura and Jensen, 2007,
Slowik et al., 2004, Gary-Ong and Myint-Lay, 2006, Igarashi and Kawamura, 2002, Lura
et al., 2007, Turcry and Loukili, 2006).
Table 2-10: Methods used for measuring shrinkage at early-age.
Shrinkage
Sample
size (m
m)
Sample
Shape
Material
Measuring
approach
Reference
100 x100 x 500 Prism Inductive displacement tra‐ nsducer &LVDTs
(Bjøntegaard et al., 2004)
100 x 100 x 400 Prism LVDTs & Embed‐ ded strain gauge
(Lee et al., 2003)
Diameter:125 Height: 250
Cylinder Embedded strain gauge
(Termkhajornkit et al., 2005)
270 x 270 x 100 Slab Vertical cast‐in bars & LVDTs
(Tazawa, 1999)
Diameter:100 Length: 375
Flexible tube
‐ LVDTs (Hansen and Jensen, 1997)
150 x 150 x 375 Beam ‐ Horizontal cast‐in bars & LVDTs
(Lokhorst, 1999)
Autogen
ous
380 x 1100 x 2000
Slab ‐ Fiber optical sensors
(Glišic and Simon, 2000)
Total 100 x 100 x
500 Prism
‐ High resolution digital camera
(Gary‐Ong and Myint‐Lay, 2006)
Autogen
ous and
Total
70 x 70 x 1000 Dog bone
Concrete
‐ LVDTs (Igarashi and Kawamura,
2002)
Chapter 2 50
Table 2-10 Cont’d: Methods used for measuring shrinkage at early-age.
Shrinkage
Sample
size (m
m)
Sample
Shape
Material
Measuring
approach
Reference
355 x 100 x 560 Slab with stress riser M
ortar
‐ Automated image analysis
(Lura et al., 2007)
Plastic
70 x 70 x 280 Prism
Concrete
‐ Laser sensors (Turcry and Loukili, 2006)
40 x 40 x 160 Prism
Mortar
‐ Laser sensors (Tazawa, 1999)
Diameter:30 Length: 300
Corrugated tube
‐ Inductive displacement transducer
(Jensen and Hansen, 1995)
Diameter:30 Length: 350
Corrugated tube
‐ Thermal comparator sensor
(Østergaard and Jensen, 2003)
Diameter:76 Length: 305 Membrane
thickness:0.64
Flexible cylindrical
latex membrane
‐ Dual‐compartment cell
(Gagné et al., 1999)
Polyurethane membrane with thickness 0.04
‐‐‐‐ ‐ Buoyancy method
(Lura and Jensen, 2007)
Autogen
ous
30 x 30 x 90 Prism
Cemen
t Paste
‐ Fiber optical sensors
(Slowik et al., 2004)
2.2.3.4 Factors Affecting Early-Age Shrinkage
The parameters that affect early-age shrinkage of concrete differ according to the driving
forces behind each shrinkage mechanism. Drying shrinkage is mainly attributed to
environmental conditions. Thus, problematic factors are those that affect the rate of
Chapter 2 51
evaporation, i.e. relative humidity, air velocity, air and concrete temperature (Esping,
2007). Additionally, the materials used in concrete have a parallel effect through
controlling the quantity and duration of bleeding and setting time (Holt and Leivo, 2004).
On the other hand, autogenous shrinkage is influenced by the type and properties of the
binder, mixture proportions, and admixtures that can refine the pore structure. Although
the early-age autogenous shrinkage is fully attributed to chemical shrinkage, it may
behave differently than chemical shrinkage with respect to some factors especially those
that affect the setting time and formation of a restraining structure (Esping, 2007).
Generally, the early-age shrinkage is affected by the following parameters:
a) The binder type, content and rate of hydration: A higher rate of hydration results in
higher autogenous and drying shrinkage due to the decreased volume of hydration
products relative to their constituents and the higher water consumption, which in turn
reduces bleeding and increases the concrete temperature (Bentur, 2003, Holt, 2005,
Tazawa, 1999, Esping, 2007, Topçu and Elgün, 2004).
b) The aggregate content: aggregates reduce shrinkage, and act as an internal
restraint. It also reduces the volume of cement paste, leading to lower chemical shrinkage
(Holt, 2001, Esping, 2007). Furthermore, light-weight aggregates with high absorption
were found to reduce autogenous shrinkage as it acts as an internal curing source (Nassif
et al., 2003, Mihashi and Leite, 2004, Tazawa, 1999, Zhutovsky et al., 2002, Duran-
Herrera et al., 2007).
c) The water content: it has a major role during early-age shrinkage through
controlling the amount of free water, and the development of the microstructure and pore
Chapter 2 52
system, which consequently affects the capillary tension and meniscus development
(autogenous shrinkage).
d) Admixtures: a shrinkage reducing admixture (SRA) can decrease the surface
tension of the capillary pore solution resulting in a reduction of the capillary tension. It
was found to reduce both autogenous and drying shrinkage strains (Bentur, 2003,
D’Ambrosia, 2002, Holt, 2005, Esping, 2007, Lura et al., 2007, Bentza et al., 2001).
About 60% reduction in the early-age unrestrained shrinkage due to an SRA dosage of 1-
2% by mass of the cement was observed within the first week after casting concrete
(D’Ambrosia, 2002). SRA has a more significant effect at low w/c as shown in Fig. 2-12.
Moreover, SRA was found to induce early expansion after the time of set, maintain a
higher relative humidity level in the concrete, and to facilitate water loss from smaller
pores, which results in reducing the concrete shrinkage as discussed by (Weiss et al.,
2008). Greater early-age shrinkage was observed with the addition of superplasticizer
(SP) as a result of improving cement dispersion, which consequently increases the rate of
hydration reactions (Holt, 2005, Esping, 2007). In addition, it was pointed out that
excessive SP dosage would delay the setting time and result in higher early-age drying
shrinkage (Holt and Leivo, 2004). Scattered data on the effect of using air entraining
admixtures on the early-age shrinkage were observed (Hammer and Fosså, 2006).
However, the general trend was an increase of the autogenous shrinkage rate before it
diverged from chemical shrinkage, and then it decreased thereafter. Conversely, air
entraining admixtures were found by others to cause a considerable decrease in early-age
shrinkage (Kronlof et al., 1995).
Chapter 2 53
-350
-300
-250
-200
-150
-100
-50
0
50
1 2 3 4 5 6 7 8Age (Days)
Free
Shr
inka
ge (m
e)
w/c=0.35w/c=0.35+ 1%SRAw/c=0.45w/c=0.45+1%SRA
Figure 2-12: Free shrinkage with different w/c and 1% shrinkage reducing
admixture [adapted after (D’Ambrosia, 2002)].
e) Pozzolanic materials: the type, fineness, and percentage of cement replacement are
the main parameters controlling the effect of pozzolanic materials on early-age shrinkage.
Silica fume was found to increase the autogenous shrinkage significantly due to refining
the pore structure of concrete (Tazawa, 1999, Wiegrink et al., 1996). Similar behaviour
was observed for ground granulated blast furnace slag (Lee et al., 2006, Lim and Wee,
2000). A high level of cement replacement by metakaolin (MK) (10-15%) was found to
reduce both autogenous and drying shrinkage at early-age. This reduction may be a result
of the dilution effect of reducing the cement content (Brooks and Megat-Johari, 2001,
Kinuthia et al., 2000). Conversely, Gleize et al. (Gleize et al., 2007) observed an increase
in autogenous shrinkage due to MK addition, which was interpreted as a result of the
heterogeneous nucleation of hydration products on the surface of MK particles. Fly ash
Chapter 2 54
was also found to reduce the autogenous shrinkage of concrete (Lee et al., 2003,
Termkhajornkit et al., 2005). In contrast, it was reported that very fine fly ash had a
similar effect to that of silica fume (Tazawa, 1999).
f) Curing conditions: the curing method, duration and temperature have a significant
effect on early-age shrinkage. Shrinkage was found to increase with increasing curing
temperature during the first 5 to 10 hours after casting, and then it decreased with
temperature increase. This may be attributed to an increase in moisture loss at high curing
temperature, leading to an increase in the development of plastic shrinkage (Zhao, 1990).
Additionally, moist curing with burlap and/or cotton mats was found to effectively reduce
the early-age shrinkage and to increase water retention compared with other curing
methods (Nassif et al., 2003, Huo and Wong, 2006). Furthermore, the longer the curing
period, the lower was the shrinkage deformation (Tazawa, 1999).
2.2.3.5 Compensating for Early-Age Shrinkage
Different methods have been developed to compensate for and/or reduce early-age
shrinkage. Table 2-11 summarizes some of these different methods (Mehta and
Monteiro, 2006, Wongtanakitcharoen and Naaman, 2007, Brooks and Megat-Johari,
2001, Bentur, 2003, D’Ambrosia, 2002, Holt, 2005, Mihashi and Leite, 2004, Tazawa,
1999, Esping, 2007, Lura et al., 2007, Duran-Herrera et al., 2007, Bentza et al., 2001,
Kronlof et al., 1995, Kinuthia et al., 2000, Sule and van Breugel, 2001, Weiss, 1999).
However, the shrinkage of concrete cannot be completely avoided; it will occur at least
due to the volume reduction resulting from the hydration of cement and water (chemical
shrinkage also called Le Chatelier's contraction).
Chapter 2 55
Table 2-11: Shrinkage compensating methods.
Method Effective parameter Action Reference Low and moder‐ ate heat cement
High C2S content & formation of ettringite
Expansion (Tazawa, 1999)
Expansive cement (Type K)
Formation of ettringite Hydration of hard‐burnt CaO.
Expansion and chemical pre‐stress level
(Mehta and Monteiro, 2006)
Shrinkage reducing admixture
Reduce surface tension of pore solution.
Reduce capillary tension
(Bentur, 2003, D’Ambrosia, 2002, Holt, 2005, Esping, 2007, Lura et al., 2007, Bentza et al.,
2001)
Fiber High elastic modulus. Resist exceeded tensile stress
(Wongtanakitcharoen and Naaman, 2007, Kronlof et al., 1995)
Internal curing
Soaked lightweight aggWater soluble, chemicals (self‐curing admixture), Smart parafine microcapsule, Superabsorbent polymer.
Store water for internal curing, increase water retention, Mitigate rapid temperature change.
(Mihashi and Leite, 2004, Esping, 2007, Duran‐Herrera et al.,
2007)
Increase w/c Free water.
Sufficient amount of water for hydration and bleeding
(Tazawa, 1999)
Reinforcement schemes
Light reinforcement bars.
Resist exceeded tensile stress
(Sule and van Breugel, 2001, Weiss, 1999)
Additive
Expansive additive, gypsum, fly ash, metakaolin, water repellent powder.
Hydration product with tendency to volume increase
(Brooks and Megat‐Johari, 2001, Tazawa, 1999, Kinuthia et al.,
2000)
2.2.4. Early-Age Creep of Concrete
Creep is a complex phenomenon, particularly at early-age, due to the rapid changes in
concrete properties (Khan, 1995) and the difficulty to clearly distinguish between
instantaneous elastic strains and early-age creep (Neville, 1996). The total creep, either
Chapter 2 56
under tensile or compressive load, is the sum of two main components: basic creep
(without external moisture loss), and drying creep (Pickett effect), which is an additional
creep due to moisture loss during loading. However, the tensile creep has greater
importance at early-age when the crack potential is to be determined (Altoubat, 2002,
Bissonnette and Pigeon, 1995).
The mechanism of early-age creep is not completely understood. Several
mechanisms have been proposed including real and apparent mechanisms. Real
mechanisms, involving the viscous flow theory, plastic flow, and seepage of gel water,
are related to cement hydration and can be considered as material properties (Bentur,
2003). Apparent mechanisms are associated with micro-cracking and stress-induced
shrinkage (Altoubat, 2002). Generally, creep is a combination of these two categories of
mechanisms. Several mathematical formulas were proposed to predicate and model early-
age creep, see (Springenschmid, 1998). However, the accuracy of these formulas in
capturing and simulating the early-age behaviour of concrete is still questionable and
need more investigation (Springenschmid, 1998).
Creep is mainly affected by the mixture composition of concrete, loading age and
duration, water migration, temperature, moisture conditions, and the stress level (Neville,
1996). The earlier the loading age, the higher the creep strain values due to the low
modulus of elasticity of concrete (Bissonnette and Pigeon, 1995, Neville, 1996).
Furthermore, no relationship between the applied stress and the resultant creep strain was
found for specimens loaded at an age of 24 hours (Østergaard et al., 2001). Increasing the
temperature was found to enhance the early-age creep rate, which is contrary to its effect
at later ages (Neville, 1996). However, high temperature is expected to reduce the early-
Chapter 2 57
age creep value as a result of accelerating the hydration process (Springenschmid, 1998).
Moreover, the early-age creep of restrained concrete was found to be inversely related to
the w/c (Altoubat, 2002).
Generally, mineral additives such as fly ash, metakaolin and slag were found to
reduce the tensile creep and relaxation of concrete at early-age (Brooks and Megat-
Johari, 2001, Pane and Hansen, 2002, Li et al., 2002), while silica fume showed an
opposite trend (Kovler et al., 1999, Bissonnette and Pigeon, 1995, Pane and Hansen,
2002, Li et al., 2002). The spherical shape and smooth surface of un-reacted silica fume
particles at early-age were suggested as a reason for this behaviour (Kovler et al., 1999).
Autogenous shrinkage was found to affect basic creep significantly, especially at
low w/c (Pane and Hansen, 2002, Li et al., 2002). Autogenous shrinkage is a material
property that is not dependent on the applied stress. Thus, it should be subtracted from
the total creep strain (see Eq. 2-7) (Lee et al., 2006). Consequently, the measured creep
can be categorized into real creep (excluding autogenous shrinkage), and apparent creep
(including autogenous shrinkage) (Fig. 2-13). One interesting observation is that SRA
was found to reduce the total tensile creep as well as the early-age shrinkage, but to a less
extent. This can be a benefit if creep is considered to offset the induced tensile stress due
to shrinkage (D’Ambrosia, 2002).
)(),(),(
),(t
ttttttJ autogenoustotal
real ′
′−′=′
σεε
Eq. 2-7
Where, is the real compliance function, )),( ttJ real ′ ,( tttotal ′ε is the total strain of a
concrete specimen measured from the basic creep test at time t caused by an amount of
Chapter 2 58
constant stress applied at age t′ , ),( ttautogenous ′ε is the autogenous shrinkage strain at time
, t σ is the total applied stress.
Figure 2-13: Apparent and real creep compliance function [after (Lee et al., 2006)].
Furthermore, the sustained mechanical loading at early-age seems to increase the
degree of hydration during early-age creep (Tamtsia et al., 2004). It was also shown that
biaxial sustained compressive load would promote hydration reactions (Liu et al., 2002).
Enhancing the hydration process involves the formation of more calcium silicate hydrate
(CSH) gel, which gradually carries load leading to stress reduction in the originally
loaded CSH gel and consequently reduced creep at early-age (Springenschmid, 1998).
Chapter 2 59
2.3. MODELS FOR SIMULATION AND ANALYSIS OF EARLY-AGE
PROPERTIES
Several complex and interacting factors affect the early-age properties of concrete, but
the hydration process is generally considered as the main indication of early-age
behaviour. Hydration of cement is a progressing system of reactions concurrent with a
nonlinear temperature rise, thermal expansion, micro-structural changes, water loss and
autogenous deformation. Therefore, developing accurate models for the early-age
behaviour of hydrated cement should be based on an accurate simulation of the cement
hydration process. Several analytical models for simulating the cement hydration process
have been developed. According to the scale at which the hydration process is described,
these models were classified into micro-scale, meso-scale, and macro-scale models
(Mihashi and Leite, 2004). In addition, a digital-image-based model using the so-called
“cellular automata” was introduced to study the hydration process, formation of
microstructure, growth of void structures, strength development, and shrinkage (Bentz
and Garboczi, 1993). However, this model was not able to simulate the hydration process
at different material scales. Therefore, a multi-scale modeling approach was introduced to
overcome this limitation, where separate models at different material scales are
developed and then connected with each other (Bentz and Garboczi, 1996). This can
better characterize for the hydration process of cement versus time.
Several other models have been introduced to predict the early-age concrete
deformation, including thermal, autogenous, drying, and creep deformations. The basic
concepts for some of these models that deal mainly with various types of early-age
deformations are reviewed hereafter.
Chapter 2 60
A thermo-dynamic model was proposed to describe the autogenous deformation
of hardening cement paste. This model was based on reaching thermo-dynamic
equilibrium in the capillary pore system, which continuously changes as a result of the
hydration progress and autogenous shrinkage. The pore water surface tension was
considered as the main reason for autogenous shrinkage (Koenders and van Breugel,
1997). However, the formula used to evaluate autogenous shrinkage is only valid for low
relative humidity (< 40%). Moreover, a macro-scale model to explain autogenous
shrinkage of cement paste was introduced (Hual et al., 1995). In this model, microscopic
stresses due to capillary depression were considered as the main driving force for
autogenous shrinkage. This model was further enhanced to estimate autogenous
shrinkage at the scale of hydrating grains (Hual et al., 1997). Likewise, a multi-scale
model for early-age autogenous shrinkage was proposed (Pichler et al., 2007). It
demonstrates the role of capillary depression and ettringite formation pressure on
autogenous shrinkage.
On the other hand, the Kelvin chain model was modified to simulate basic creep
in early-age concrete (De Schutter, 1999). The modified Kelvin chain model units have
variable parameters that vary depending on the degree of hydration. As a result, the
model was able to simulate the visco-elastic behaviour of early-age concrete and a good
agreement with experimental results was observed. More recently, a hygro-thermo-
chemo-mechanical model for early-age creep was introduced. In this model, creep was
described based on the micro-pre-stress-solidification theory, which was modified to take
into account the effect of relative humidity on the cement hydration rate and the
associated hygro-thermal phenomena (Gawin et al., 2006). The good agreement with
Chapter 2 61
experimental data indicates a potential of this model to analyze creep phenomena at
different ages and environmental conditions.
Generally, a challenge remains as to how to deal with the nonlinear ageing
behaviour of concrete, and how to characterize the behaviour of its constituents at the
microscopic and nano-scale levels under field mechanical and environmental conditions.
Generally, a smart model which can predicate and demonstrate the interaction effects
between different types of deformations (i.e. shrinkage and creep) during early-age of
cement-based materials still remain to be developed.
2.4. CONCLUDING REMARKS
It is widely accepted that the early-age period of concrete has a substantial influence on
the long-term service life performance of concrete structures. During early-age,
continuous and rapid changes in the concrete microstructure and constituents proportions
(un-hydrated and hydrated cement, water content, etc.) take place. Consequently, the
early-age characteristics of concrete, including thermal, mechanical and deformation
properties, can have different development rates, relationships and values compared to
that of the later age properties. Moreover, early-age properties are generally more
sensitive to the curing regime (i.e. temperature and relative humidity), early-age loading
rate and value, and the use of chemical and mineral admixtures. Several models for the
early-age behaviour of concrete have been proposed. It appears that simulating the
cement hydration process on a multi-scale level is a useful modeling method for
predicting the early-age behaviour as it captures the microscopically ageing cement-based
materials behaviour.
Chapter 2 62
For a better understanding and control of the early-age properties of cement-based
materials, especially new and emerging materials such as ultra-high performance
concrete, the following remarks are noteworthy:
1) Optimization of the binder composition (e.g. use of SCMs) can allow achieving
superior early-age performance along with avoiding various divergent effects on
early-age properties.
2) There is a need for improving measuring techniques to capture the early-age
properties precisely as early as possible and for developing standard testing methods
to avoid conflicting data of various non-standardized test methods.
3) Early-age properties of concrete should be investigated under simulated field-like
conditions to capture the real behaviour of cement-based materials and to develop
more realistic predictive models and relationships.
4) Relationships between early-age concrete properties are greatly affected by the
quality of concrete, which is influenced by several factors including the mixture
proportions, constituent quality, ageing, etc. Hence, existing formulas for these
properties in the literature should be calibrated to the actual concrete mixture and
exposure conditions.
5) Mitigating the cracking of concrete should be studied in a holistic manner including
thermal strains, drying and autogenous shrinkage.
6) Coupling new modeling tools such as artificial intelligence with existing empirical
models may facilitate capturing the interactions between different early-age
properties of concrete.
Chapter 2 63
2.5. REFERENCES
Abel, J. and Hover, K., (1998), “Effect of water/cement ratio on the early-age tensile strength of concrete,” Transportation Research Record: Journal of the Transportation Research Board, No. 1610, pp. 33-38.
ACI Committee 318, (2005). Building Code Requirements for Structural Concrete (ACI
318-05) and Commentary (ACI 318R-05). 2nd ed., American Concrete Institute, Michigan.
Aitcin, P.C., (2003), “The durability characteristics of high performance concrete: a
review,” Cement and Concrete Composites, Vol. 25, No. 4–5, 409-420. Alshamsi, A.M., (1997), “Microsilica and ground granulated blast furnace slag effects on
hydration temperature,” Cement and Concrete Research, Vol. 27, No. 12, pp. 1851-1859.
Altoubat, S.A., (2002), “Early-age stresses and creep-shrinkage interaction of restrained
concrete,” PhD Thesis, University of Illinois at Urbana-Champaign, United States.
ASTM American Society for Testing and Materials, (2007). Standard Specification for
Portland Cement. ASTM International, West Conshohocken, ASTM C150 - 07. ASTM American Society for Testing and Materials, (1994). Standard Test Method for
Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression. ASTM International, West Conshohocken, ASTM C 469-1994.
Atrushi, D.S., (2003), “Tensile and compressive creep of early-age concrete: testing and
modeling,” PhD Thesis, The Norwegian University of Science and Technology, Norway.
Bentur, A. (ed.), (2003). Early-age Cracking in Cementitious Systems. RILEM Technical
Committee 181-EAS: Early-age Shrinkage Induced Stresses and Cracking in Cementitious Systems. RILEM, Report 25, 335 p.
Bentz, D.P. and Garboczi, E.J., (1993), “Digital-image-based computer modelling of
cement-based materials,” In: Digital Image Processing: Techniques and Application in Civil Engineering (Frost, J. D. & Wright, J. R. (ed.)). ASCE, New York, pp. 44-51.
Bentz, D.P. and Garboczi, E.J., (1996), “Digital-image-based computer modelling and
visualization of cement-based materials,” In: Transportation Research Record: Emerging technology in Geotechnical Engineering, No. 1526, National Academy Press, Washington DC., pp. 129-134.
Chapter 2 64
Bentz, D.P., (2007), “Transient plane source measurements of the thermal properties of hydrating cement pastes,” Materials and Structures, Vol. 40, No. 10, pp. 1073-1080.
Bentza, D.P., Geiker, M.R. and Hansen, K.K., (2001), “Shrinkage-reducing admixtures
and early-age desiccation in cement pastes and mortars,” Cement and Concrete Research, Vol. 31, No. 7, pp. 1075-1085.
Bissonnette, B. and Pigeon, M., (1995), “Tensile creep at early-ages of ordinary, silica
fume and fiber-reinforced concretes,” Cement and Concrete Research, Vol. 25, No. 5, pp. 1075-1085.
Bjøntegaard, Ø. and Sellevold, E.J., (2001), “Interaction between thermal dilation and
autogenous deformation in high-performance concrete,” Materials and Structures, Vol. 34, No. 5, pp. 266-272.
Bjøntegaard, Ø., Hammer, T.A. and Sellevold, E. J. (2004), “On the measurement of free
deformation of early-age cement paste and concrete,” Cement and Concrete Composites, Vol. 26, No. 5, pp. 427-435.
Brooks, J.J. and Megat-Johari, M.A., (2001), “Effect of metakaolin on creep and
shrinkage of concrete,” Cement and Concrete Composite, Vol. 23, No. 6, pp. 495-502.
Brown, T.D. and Javaid, M.Y., (1970), “The thermal conductivity of fresh concrete,”
Materials and Structures, Vol. 3, No. 6, pp. 411-416. BSI British Standards. (1983). Method for Determination of Static Modulus of Elasticity
in Compression. BSI, London, BS 1881: Part 121. Byfors, J. (1980), “Plain Concrete at Early-ages,” Swedish Cement and Concrete
Institute, Stockholm, Technical Report, 345 p. CEB-FIP Model Code for Concrete Structures 1990. (1991), Evaluation of the Time
Dependent Behaviour of Concrete. Bulletin d’ Information No. 199, Comité Europeen Du Béton / Fédération Internatonale De La Précontrainte, Lausanne.
Cusson, D. and Hoogeveen, T.J., (2006), “Measuring early-age coefficient of thermal
expansion in high-performance concrete,” In: Proceeding of the International RILEM Conference on Volume Changes of Hardening Concrete: Testing and Mitigation. Denmark RILEM Publications SARL, Lyngby, pp. 321-330.
D’Ambrosia, M., (2002), “Early-age Tensile Creep and Shrinkage of Concrete with
Shrinkage Reducing Admixtures,” Master Thesis, University of Illinois at Urbana-Champaign, United States.
Chapter 2 65
D’Ambrosia, M., Altoubat, S.A., Park, C. and Lange, D.A., (2001), “Early-age tensile creep and shrinkage of concrete with shrinkage reducing admixtures. In Creep, Shrinkage and Durability Mechanics of Concrete and other Quasi-Brittle Materials (Ulm, F. J., Bazant, Z. P. & Wittman, F. H. (ed.)). Elsevier, Cambridge MA, pp. 645-651.
De Schutter, G. and Taerwe, L., (1995), “Specific heat and thermal diffusivity of
hardening concrete,” Magazine of Concrete Research, Vol. 47, No. 172, pp. 203-208.
De Schutter, G. and Taerwe, L., (1996), “Degree of hydration-based description of
mechanical properties of early-age concrete,” Materials and Structures, Vol. 29, No. 6, pp. 335-344.
De Schutter, G., (1999), “Degree of hydration based Kelvin model for the basic creep of
early-age concrete,” Materials and Structures, Vol. 32, No. 4, pp. 260-265. Downie, B., (2005), “Effect of moisture and temperature on the mechanical properties of
concrete,” Master Thesis, West Virginia University, West Virginia, United States. Duran-Herrera, A., Aitcin, P.C. and Petrov, N., (2007), “Effect of saturated lightweight
sand on shrinkage in 0.35 w/b concrete,” ACI Materials Journal, Vol. 104, No. 1, pp. 48-52.
Esping, O., (2007), “Early-age properties of self-compacting concrete - Effects of fine
aggregate and limestone filler,” PhD Thesis, Department of Civil and Environmental Engineering, Chalmers University of Technology, Sweden.
Fu, X. and Chung, D.D.L., (1997), “Effects of silica fume, latex, methycellulose, and
carbon fibers on the thermal conductivity and specific heat of cement paste,” Cement and Concrete Research, Vol. 27, No. 12, pp. 1799-1804.
Gagné, R., Aouad, I., Shen, J. and Poulin, C., (1999), “Development of a new
experimental technique for the study of the autogenous shrinkage of cement paste,” Materials and Structures, Vol. 32, No. 9, pp. 635-642.
Gardner, N.J. and Zhao, J.W., (1993), “Creep and shrinkage revisited,” ACI Materials
Journal, Vol. 90, No. 3, pp. 236-246. Gary-Ong, K.C. and Myint-Lay, K., (2006), “Application of image analysis to monitor
very early-age shrinkage,” ACI Materials Journal, Vol. 103, No. 3, pp. 169-176. Gawin, D., Pesavento, F. and Schrefler, B.A., (2006), “Hygro-thermo-chemo-mechanical
modelling of concrete at early ages and beyond Part II: Shrinkage and creep of concrete,” International Journal for Numerical Methods in Engineering, Vol. 67, No. 3, pp. 332-363.
Chapter 2 66
Ghali, A. and Favre, R. (ed.) (1994). Concrete Structures: Stresses and Deformations, 3rd
edn., Taylor, London. Gleize, P.J.P., Cyr, M. and Escadeillas, G., (2007), “Effects of metakaolin on autogenous
shrinkage of cement pastes,” Cement and Concrete Composites, Vol. 29, No. 2, pp. 80-87.
Glišic, B. and Simon, N., (2002), “Monitoring of concrete at very early-age using stiff
SOFO sensor,” Cement and Concrete Composites, Vol. 22, No. 2, pp. 115-119. Goldsmid, H. J. (ed.) (1965). The Thermal Properties of Solids. 1th edn., Routledge and
Kegan Paul, London. Hammer, T.A. and Fosså, K.T., (2006), “Influence of entrained air voids on pore water
pressure and volume change of concrete before and during setting,” Materials and Structures, Vol. 39, No. 9, pp. 801-808.
Hammer, T. A., Fosså, K. T. and Bjøntegaard, Ø., (2007), “Cracking tendency of HSC:
Tensile strength and self generated stress in the period of setting and early hardening,” Materials and Structures, Vol. 40, No. 3, pp. 319-324.
Hansen, K.K. and Jensen, O.M., (1997), “Equipment for measuring autogenous RH-
change and autogenous deformation in cement paste and concrete,” In: Proceedings of the 1st International Research Seminar on Self-desiccation and its Importance in Concrete Technology (Persson, B. and Fagerlund, F. (ed.)). Lund, pp. 27-30.
Holt, E. and Leivo, M., (2004), “Cracking risks associated with early-age shrinkage,”
Cement and Concrete Composite, Vol. 26, No. 5, pp. 521-530. Holt, E.E., (2001a). “Early-age autogenous shrinkage of concrete,” PhD Thesis,
University of Washington, United States. Holt, E.E., (2001b), “Early age autogenous shrinkage of concrete,” Espoo 2001.
Technical Research Centre of Finland, VTT Publications 446, 184 p. Holt, E.E., (2005), “Contribution of mixture design to chemical and autogenous
shrinkage of concrete at early-ages,” Cement and Concrete Research, Vol. 35, No. 3, pp. 464-472.
Hual, C., Ehrlacher, A. and Acker P., (1997), “Analyses and models of the autogenous
shrinkage of hardening cement paste: II. Modelling at scale of hydrating grains,” Cement and Concrete Research, Vol. 27, No. 2, pp. 245-258.
Chapter 2 67
Hual, C., Ehrlacher, A. and Acker, P., (1995), “Analyses and models of the autogenous shrinkage of hardening cement paste: I. Modeling at microscopic scale,” Cement and Concrete Research, Vol. 25, No. 7, pp. 1457-1468.
Huo, X.S. and Wong, L.U., (2006), “Experimental study of early-age behavior of high-
performance concrete deck slabs under different curing methods,” Construction and Building Materials, Vol. 20, No. 10, pp. 1049-1056.
Igarashi, S. and Kawamura, M., (2002), “Effects of microstructure on restrained
autogenous shrinkage behavior in high strength concretes at early-ages,” Materials and Structures, Vol. 35, No. 2, pp. 80-84.
Jensen, O.M. and Hansen, P.F., (1995), “A dilatometer for measuring autogenous
deformation in hardening Portland cement paste,” Materials and Structures, Vol. 8, No. 7, pp. 406-409.
Kada, H., Lachemi, M., Petrov, N., Bonneau, O. and Aïtcin, P.C., (2002), “Determination
of the coefficient of thermal expansion of high-performance concrete from initial setting,” Materials and Structures, Vol. 35, No. 1, pp. 35-41.
Kahouadji, A., Clastres, P. and Debicki, G., (1997), “Early-age compressive strength
prediction of concrete - application on a construction site,” Construction and Building Materials, Vol. 11, No. 7-8, pp. 431-436.
Kasai, Y., Yokoyama, K. and Matsui, T., (1974), “Tensile properties of early-age
concrete,” In: Proceedings of the 1974 Symposium on Mechanical Behaviour of Materials. Kyoto, pp. 433-441.
Khan, A.A., (1995), “Concrete properties and thermal stress analysis of members at
early-ages,” PhD Thesis, McGrill University, Canada. Kinuthia, J.M., Wild, S., Sabir, B.B. and Bai, J., (2000), “Self-compensating autogenous
shrinkage in Portland cement-metakaolin-fly ash pastes,” Advances in cement research, Vol. 12, No. 1, pp. 35-43.
Koenders, E.A.B. and van Breugel, K., (1997), “Numerical modeling of autogenous
shrinkage of hardening cement paste,” Cement and Concrete Research, Vol. 27, No. 10, pp. 1489-1499.
Kook-Han, K., Sang-Eun, J., Jin-Keun, K. and Sungchul, Y., (2003), “An experimental
study on the thermal conductivity of concrete,” Cement and Concrete Research, Vol. 33, No. 3, pp. 363-371.
Kosmatka, S.H., Kerkhoff, B., Panarese, W.C., Macleod, N.F. and McGrath, R.J. (ed.)
(2002), “Design and Control of Concrete Mixtures,” Cement Association of Canada, 7th Edition.
Chapter 2 68
Kovler, K., Igarashi, S. and Bentur, A., (1999), “Tensile creep behavior of high-strength concretes at early-ages,” Materials and Structuresl, Vol. 32, No. 5, pp. 383-387.
Kronlof, A., Leivo, M. and Sipari, P., (1995), “Experimental study on the basic
phenomena of shrinkage and cracking of fresh mortar,” Cement and Concrete Research, Vol. 25, No. 8, pp. 1747-1754.
Lea, F.M., (2004). Lea's Chemistry of Cement and Concrete, 4th Edition, (ed.) P.C.
Hewlett, Elsevier Ltd., Oxford, UK, 1057 p. Lee, K.M., Lee, H.K., Lee, S.H., Kim, G.Y., (2006), “Autogenous shrinkage of concrete
containing granulated blast-furnace slag,” Cement and Concrete Research, Vol. 36, No. 7, pp. 1279-1285.
Lee, K., Lee, K.M. and Kim, B.K., (2003), “Autogenous shrinkage of high-performance
concrete containing fly ash,” Magazine of Concrete Research, Vol. 55, No. 6, pp. 507-515.
Lee, Y., Yi, S.T., Kim, M.S. and Kim, J.K., (2006), “Evaluation of basic creep model
with respect to autogenous shrinkage,” Cement and Concrete Research, Vol. 36, No. 7, pp. 1268-1278.
Li, H., Wee, T.H., Wong, S.F., (2002), “Early-age creep and shrinkage of blended cement
concrete,” ACI Materials Journal, Vol. 99, No. 1, pp. 3-10. Lim, S.N. and Wee, T.H., (2000), “Autogenous shrinkage of ground-granulated blast-
furnace slag concrete,” ACI Materials Journal, Vol. 97, No. 5, pp. 587-593. Liu, G.T., Gao, H., Chen, F.Q., (2002), “Microstudy on creep of concrete at early-age
under biaxial compression,” Cement and Concrete Research, Vol. 32, No. 12, pp. 1865-1870.
Lokhorst, S.J., (1999), “Deformational behavior of concrete influenced by hydration-
related changes of the microstructures,” Delft University of Technology, Netherlands, Research Report 25.5-99-5.
Lura, P. and Jensen, O.M., (2007), “Measuring techniques for autogenous strain of
cement paste,” Materials and Structures, Vol. 40, No. 4, pp. 431-440. Lura, P., Pease, B., Mazzotta, G., Rajabipour, F. and Weiss, J., (2007), “Influence of
shrinkage-reducing admixtures on the development of plastic shrinkage cracks,” ACI Materials Journal, Vol. 104, No. 2, pp.187-194.
McCullough, B.F. and Rasmussen, R.O., (1999), “Fast track paving: concrete
temperature control and traffic opening criteria for bonded concrete overlays,”
Chapter 2 69
Federal Highway Administration, Washington D.C., Report FHWA RD 98-167, 191 p.
Mehta, P.K. and Monteiro, P.J.M. (ed.), (2006). Concrete: Microstructure, Properties, and
Materials. 3rd edn. McGraw-Hill, New York. Mesbah, H.A., Lachemi, M. and Aïtcin, P.C., (2002), “Determination of elastic properties
of high-performance concrete at early-age,” ACI Materials Journal, Vol. 99, No. 1, pp. 37-41.
Mihashi, H. and Leite, J.P., (2004), “State-of-art Report on Control of Cracking in Early-
age Concrete,” Journal of Advanced Concrete Technology, Vol. 2, No. 2, pp. 141-154
Mounanga, P., (2004), “Étude Expérimentale du Comportement des Pâtes de Ciment au
Très Jeune Age: Hydratation, Retraits, Propriétés Thermo-physiques,” PhD Thesis, University of NANTES, France.
Myers, J.J., (1999). How to achieve a higher modulus of elasticity. Publication – HPC
Bridge Views, FHWA, Sponsored, NCBC Co-Sponsored Newsletter, No. 5, pp. 1-3.
Nassif, H.H., Suksawang, N. and Mohammed, M., (2003), “Effect of curing method on
early–age drying shrinkage of high-performance concrete,” Transportation Research Record: Journal of the Transportation Research Board, No. 1834, TRB, National Research Council, Washington, D.C., pp. 48-58.
Neville, A.M. (ed.), (1996). Properties of Concrete. 4th Edition. Wiley, New York. Oluokun, F.A., (1989), “Investigation of physical properties of concrete at early-ages,”
PhD Thesis, The University of Tennessee, United States. Oluokun, F.A., Burdette, E.G. and Deatherage, J.H., (1991), “Elastic modulus, Poisson's
ratio, and compressive strength relationships at early-ages,” ACI Materials Journal, Vol. 88, No. 1, pp. 3-10.
Oluokun, F.A., Burdette, E.G. and Deatherage, J.H., (1990), “Early-age concrete:
Strength prediction by maturity – another look,” ACI Materials Journal, Vol. 87, No. 6, pp. 565-572.
Oluokun, F.A., Burdette, E.G. and Deatherage, J.H., (1991), “Splitting tensile strength
and compressive strength relationships at early-ages,” ACI Materials Journal, Vol. 88, No. 2, pp. 115-121.
Chapter 2 70
Østergaard, L., (2003), “Early-age Fracture Mechanics and Cracking of Concrete, Experiments and Modeling,” PhD Thesis, Technical University of Denmark, Denmark.
Østergaard, L., Lange, D.A., Altoubat, S.A. and Stang, H., (2001), “Tensile basic creep of early-age concrete under constant load,” Cement and Concrete Research, Vol. 31, No. 12, pp. 1895-1899.
Østergaard, T. and Jensen, O.M., (2003), “A thermal comparator sensor for measuring
autogenous deformation in hardening Portland cement paste,” Materials and Structures, Vol. 36, No. 10, pp. 661-665.
Pane, I. and Hansen, W., (2002), “Early-age creep and stress relaxation of concrete
containing blended cements,” Materials and Structures, Vol. 35, No. 2, pp. 92-96. Pane, I. and Hansen, W., (2005), “Investigation of blended cement hydration by
isothermal calorimetry and thermal analysis,” Cement and Concrete Research, Vol. 35, No. 6, pp. 1155-1164.
Pichler, C., Lackner, R. and Mang, H.A., (2007), “A multiscale micromechanics model
for the autogenous-shrinkage deformation of early-age cement-based materials,” Engineering Fracture Mechanics, Vol. 74, No. 1-2, pp. 34-58.
Prusinski, J.R. (2006), “Chapter 44: Slag as a cementitious material,” In Significance of
Tests and Properties of Concrete and Concrete-making Materials (Lamond, J. F. and Pielert, J. H. (ed.)). ASTM International, West Conshohocken, pp. 512-531.
Raphael, J.M., (1984), “Tensile strength of concrete,” ACI Materials Journal, Vol. 81,
No. 2, pp.158-165. RILEM Commission 42-CEA, (1981), “Properties of set concrete at early-ages,”
Materials and Structures, Vol. 14, No. 84, pp. 399-460. RILEM TC119-TCE., (1997), “Avoidance of Thermal Cracking in Concrete at Early
Ages (Recommendation: TCE1: Adiabatic and Semi-adiabatic Calorimetry to Determine the Temperature Increase in Concrete Due to Hydration Heat of the Cement),” Materials and Structures Journal, Vol. 30, No. 202, pp. 451-464.
Schindler, A.K., (2004), “Effect of temperature on hydration of cementitious materials,”
ACI Materials Journal, Vol. 101, No. 1, pp. 72-81. Sellevold, E.J. and BjØntegaard, Ø., (2006), “Coefficient of thermal expansion of cement
paste and concrete: mechanisms of moisture interaction,” Materials and Structures Journal, Vol. 39, No. 9, pp. 809-815.
Chapter 2 71
Shan-bin, M. and Zhao-jia, Z., (2002), “The early strength of slag cements with addition of hydrate microcrystals,” Journal of Wuhan University of Technology-Material Science Edition, Vol. 17, No. 2, pp. 83-85.
Slowik, V., Schlattner, E. and Klink, T., (2004), “Experimental investigation into early-
age shrinkage of cement paste by using fiber Bragg gratings,” Cement and Concrete Composites, Vol. 26, No. 5, pp. 473-480.
Snelson, D., Wild, S. and O'Farrell, M., (2008), “Heat of hydration of portland cement-
metakaolin-fly ash (PC–MK–PFA) blends,” Cement and Concrete Research, Vol. 38, No. 6, pp. 832-840.
Springenschmid, R. (ed.) (1998), “Prevention of Thermal Cracking in Concrete at Early
Ages,” RILEM Technical Committee 119, Avoidance of Thermal Cracking in Concrete at Early Ages. RILEM, Report 15, 348 p.
Sule, M. and van Breugel, K., (2001), “Cracking behavior of reinforced concrete
subjected to early-age shrinkage,” Materials and Structures Journal, Vol. 34, No. 5, pp. 284-292.
Swaddiwudhipong, S., Lu, H.R. and Wee, T.H., (2003), “Direct tension test and tensile
capacity of concrete at early-age,” Cement and Concrete Research, Vol. 33, No. 12, pp. 2077-2084.
Tamtsia, B.T., Beaudoin, J.J. and Marchand, J., (2004), “The early-age short-term creep
of hardening cement paste: load-induced hydration effects,” Cement and Concrete Composites, Vol. 26, No. 5, pp. 481-489.
Tazawa, E. and Miyazawa, S., (1995), “Experimental study on mechanism of autogenous
shrinkage of concrete,” Cement and Concrete Research, Vol. 24, No. 8, pp. 1633-1638.
Tazawa, E. (ed.) (1999). Autogenous Shrinkage of Concrete. Proceedings of the
International Workshop organized by the Japan Concrete Institute, Taylor, New York.
Termkhajornkit, P., Nawa, T., Nakai, M. and Saito, T., (2005), “Effect of fly ash on
autogenous shrinkage,” Cement and Concrete Research, Vol. 35, No. 3, pp. 473-482.
Tia, M., Liu, Y. and Brown, D. (2005), “Modulus of elasticity, creep and shrinkage of
concrete,” Department of Civil and Coastal Engineering, University of Florida, Florida, United States, Report U.F. Project No. 49104504973-12, 165 p.
Topçu, İ.B. and Elgün, V.B., (2004), “Influence of concrete properties on bleeding and
evaporation,” Cement and Concrete Research, Vol. 34, No. 2, pp. 275-281.
Chapter 2 72
Turcry, P. and Loukili, A., (2006), “Evaluation of plastic shrinkage cracking of self-
consolidating concrete,” ACI Materials Journal, Vol. 103, No. 4, pp. 272-279. Uysal, H., Demirbog, R., Sahin, R. and Gűl, R., (2004), “The effects of different cement
dosages, slumps, and pumice aggregate ratios on the thermal conductivity and density of concrete,” Cement and Concrete Research, Vol. 34, No. 5, pp. 845-848.
Viviani, M., Glisic, B., Smith, I.F.C., (2006), “System for monitoring the evolution of the
thermal expansion coefficient and autogenous deformation of hardening materials,” Smart Materials and Structures, Vol. 15, No. 6, pp. N137-N146.
Weiss, J., Lura, P., Rajabipour, F. and Sant, G., (2008), “Performance of Shrinkage-
Reducing Admixtures at Different Humidities and at Early Ages,” ACI Materials Journal, Vol. 105, No. 5, pp. 478-486.
Weiss, W.J., (1999), “Prediction of Early-age Shrinkage Cracking in Concrete Elements,”
PhD Thesis, Northwestern University, United States. Wiegrink, K., Marikunte, S. and Shah, S.P., (1996), “Shrinkage cracking of high-strength
concrete,” ACI Materials Journal, Vol. 93, No. 5, pp. 1-8. Wongtanakitcharoen, T. and Naaman, A.E., (2007), “Unrestrained early-age shrinkage of
concrete with polypropylene, PVA, and carbon fibers,” Materials and Structures, Vol. 40, No. 3, pp. 289-300.
Xu, Y. and Chung, D.D.L., (1999), “Increasing the specific heat of cement paste by
admixture surface treatments,” Cement and Concrete Research, Vol. 29, No. 7, pp. 1117-1121.
Zhao, J., (1990). “Mechanical properties of concrete at early-ages,” Master Thesis,
University of Ottawa, Canada. Zhutovsky, S., Kovler, K. and Bentur, A., (2002), “Efficiency of lightweight aggregates
for internal curing of high strength concrete to eliminate autogenous shrinkage,” Materials and Structures, Vol. 35, No. 2, pp. 97-101.
Chapter 3 73
CHAPTER THREE
INTERACTION MECHANSIMS OF EARLY-AGE SHRINKAGE OF UHPC UNDER SIMULATED FIELD-LIKE
DRYING CONDITIONS*
This chapter provides a fundamental investigation for the evaluation of different early-
age shrinkage of UHPC and their interaction mechanisms under field-like conditions.
UHPC specimens were exposed to different simulated field-like conditions including
temperatures, namely, 10, 20 and 40°C and relative humidity (RH) ranging from 40 to
80%.
3.1. INTRODUCTION
Ultra high-performance concrete (UHPC) represents a leap development in concrete
technology. Its high strength and enhanced durability make it the ultimate building
material for the construction, strengthening and rehabilitation of bridges and other
transportation infrastructure (Tang, 2004). However, UHPC is affected by high self-
desiccation and autogenous shrinkage due its low water-to-binder ratio (w/b) (Ichinomiya
et al., 2005). Moreover, exposure to drying conditions and moisture loss during early-
ages are of particular concern in thin applications of UHPC, such as in slabs on grade,
repairs and overlays, where the exposed surface area per unit volume of material is high.
Contrary to conventional drying, self-desiccation is an internal drying (without
mass loss) of capillary pores as a result of water consumption by the hydration of cement
*A version of this chapter was published online in Materials and Structures. Some parts of this chapter were also published in the Eighth International Conference on Short and Medium Span Bridges, Niagara Falls, Ontario, Canada, (2010).
Chapter 3 74
(Aïtcin, 1999). Moreover, the hydration of cement results in chemical shrinkage, which is
the driving force for self-desiccation shrinkage (Jensen and Hansen, 1999). Therefore,
self-desiccation shrinkage can be evaluated based on the progress of hydration reactions
(i.e. increase of the chemically bound water) (Yang et al., 2005, Tazawa, 1999).
Nevertheless, drying and self-desiccation shrinkage occur with similar phenomena,
especially for low w/b concrete. Since low w/b concrete typically has low bleeding, the
evaporation from its surface at low RH conditions will extract water from the internal
mass, causing moisture loss and menisci formation (Holt, 2001). This process will extend
until equilibrium is achieved between the capillary pore pressure and the vapour pressure
above the menisci (Kovler and Zhutovsky, 2006). Generally, more evaporation during
early stages will result in more shrinkage. On the other hand, localized water loss due to
hydration process will result in the formation of internal water menisci. As hydration
proceeds, the menisci radius will decrease, leading to higher capillary stresses, and
consequently to higher self-desiccation shrinkage (Holt, 2005, Jensen and Hansen, 1996).
In concrete exposed to field conditions during early-ages, water evaporates from
the surface along with self-desiccation due to the progress of hydration reactions (Yuasa
et al., 1999). Therefore, a better understanding of the combined effects of drying and
autogenous shrinkage, and their interaction mechanisms under field-like conditions is
needed. Furthermore, an accurate evaluation of the associated deformations is useful to
seek best solutions for reducing early-age deformations and the associated cracking, thus
leading to stronger, more durable and maintenance free structures.
Substantial research has focused on evaluating the autogenous shrinkage of
different concrete mixtures under sealed conditions, while limited research has explored
Chapter 3 75
the influence of drying conditions and moisture loss on the autogenous shrinkage
behaviour during early-age. In addition, only a few studies have been conducted in
environments other than 20°C and 50% RH. The present study explores the combined
effects of drying and autogenous shrinkage in UHPC and their mutual interactions under
different drying conditions.
3.2. RESEARCH SIGNIFICANCE
Several studies have been conducted to quantify autogenous strains in UHPC, either
under sealed condition without accounting for the moisture exchange between concrete
and its surrounding, or assuming that autogenous and drying strains can be superimposed.
This study adopts a more fundamental approach based on the progress of hydration
reactions in an attempt to capture the effects of ambient conditions, including temperature
and RH, on the development of autogenous strains in UHPC. The results should have
important implications in better understanding for the evolution of early-age
deformations under field-like conditions.
3.3. METHODOLOGY
To clarify the relationship between autogenous strains under sealed and drying
conditions, this relationship will be evaluated based on the superposition principle and
the degree of hydration method. The degree of hydration method will be applied in this
study as follows: First, the progress of hydration under different curing conditions will be
quantitatively evaluated based on the amount of chemically bound water (BW).
Simultaneously, autogenous and total strains under identical curing conditions will be
Chapter 3 76
measured. Subsequently, a relationship between autogenous strains under sealed
conditions and BW will be established. Finally, the contribution of autogenous strains to
total strains under drying conditions will be evaluated.
3.4. EXPERIMENTAL PROGRAM
3.4.1. Materials and Mixture Proportions
An ordinary portland cement (OPC) was used. Silica fume (SF) was added in a dry
powder form as partial replacement for cement. The chemical and physical properties of
the various binders are listed in Table 3-1. The high SF content was chosen within the
optimum SF content recommended by (Youhua, 2000, Schachinger et al., 2004) and used
in a wide range of UHPC applications to achieve the required mechanical properties
(Graybeal and Hartmann, 2005, Katrin et al., 2006). According to the suggestions of
(Cheyrezy et al., 1995), coarse aggregate was not used. Quartz sand having a particle size
in the range of 0.1 to 0.8 mm was used. A polycarboxylate-based high-range water-
reducing admixture (HRWRA) was added at a rate of 3% by mass of cement. Water from
the HRWRA was included in the specified water to cement ratio (w/c). The mixture
compositions of the control UHPC mixtures with a target compressive strength of 150
MPa are shown in Table 3-2.
3.4.2. Environmental Conditions
To understand the mechanisms that govern UHPC shrinkage, it is necessary to
understand the real effects that exist in structural elements cast in-situ as well as in
precast elements subjected to heat curing. Heat curing of UHPC usually consists of steam
Chapter 3 77
curing at 90°C for two days, which is not always easy to apply in the precast industry or
in-situ. Therefore, different curing conditions (i.e. temperature and RH) were applied in
the present study in order to investigate the influence of in-situ environmental conditions
and normal temperature curing regimes during early-age on UHPC deformations (Table
3-3). At cold curing conditions, maintaining the RH as low as 40% inside the walk-in
environmental chamber was not feasible. Therefore, this condition was excluded from the
experimental program.
Table 3-1: Chemical and physical properties of cement and Silica fume
OPC Silica Fume
SiO2 (%) 19.8 94.0 CaO (%) 63.2 0.4
Al2O3 (%) 5.0 0.1
Fe2O3 (%) 2.4 0.1
MgO (%) 3.3 0.4
K2O (%) 1.2 0.9
SO3 (%) 3.0 1.3
Na2O (%) 0.1 0.1
TiO2 (%) 0.3 0.3
CaCO3 (%) ‐‐ ‐‐
Loss on ignition (%) 2.5 4.7
Specific surface area (m2/kg) 410 19530 Specific gravity 3.17 2.12
C3S 61 ‐‐
C2S 11 ‐‐
C3A 9 ‐‐
C4AF 7 ‐‐
Chapter 3 78
Table 3-2: Composition of control mixture.
Material (mass/cement mass)
Cement 1.00
Silica fume 0.30
Quartz sand (0.1‐0.5 mm) 0.43
Quartz sand (0.3‐0.8 mm) 1.53
water 0.22 / 0.25
HRWRA 0.03
Table 3-3: Simulated environmental conditions.
Curing condition Temperature (°C) Ambient humidity (%)
Cold 10 60 / 80
Normal 20 40 / 60 / 80
Hot 40 40 / 60 / 80
3.4.3. Preparation of Test Specimens and Testing Procedures
Cubic specimens (50×50×50 mm) were used to determine the compressive strength at 1,
3, 7 and 28 days according to ASTM C109/C109M-08 (Standard Test Method for
Compressive Strength of Hydraulic Cement Mortars (Using 2-in. or [50-mm] Cube
Specimens)). Prismatic specimens (25×25×280 mm) were used to measure autogenous
and total strains. Identical size specimens were used to measure the mass loss, coefficient
of thermal expansion (CTE) and thermo-gravimetric analysis (TGA) tests in order to
dispel the effect of the specimen size on the results. Microanalysis was conducted on a
chip of specimen from selected UHPC mixtures. The fracture surfaces were examined
Chapter 3 79
using scanning electron microscopy coupled with energy dispersive X-ray analysis
(SEM/EDX) using a Hitatchi S-4500 Field Emission SEM.
All UHPC specimens were taken from a single batch of the tested mixture.
Specimens were cast in layers and compacted on a vibrating table. After casting,
specimens were maintained at ambient temperature (i.e. 20 ±1 °C) and covered with
polyethylene sheets until demolding to avoid any moisture loss. All specimens were
demolded at the final setting time and initial readings were taken before moving
specimens to the pre-determined curing conditions inside the walk-in environmental
chamber (Note: strains will be used hereafter to account for both shrinkage and
expansion deformations).
3.4.3.1 Chemically Bound Water and Degree of Hydration
Thermo-gravimetric and derivative thermo-gravimetric analyses (TGA/DTG) were used
to determine the evolution of the BW content during hydration. This indirect method has
been commonly used (e.g. Loukili et al., 1999, Mounanga et al., 2004) to quantify the
degree of hydration. Since only one binder composition was used, a linear correlation
between the amount of BW and the degree of hydration was assumed, in agreement with
previous studies (Loukili et al., 1999, Mounanga et al., 2004, Kjellsen and Detwiler,
1992, Lam et al., 2000, Pane and Hansen, 2005). At the specific testing age, small pieces
were sliced from (25×25×280 mm) specimens and submerged in an isopropanol solvent
to stop hydration. They were subsequently dried using a desiccator until a constant mass
was achieved. Immediately before testing, the dried samples were ground and sieved on a
No. 350 sieve. The tested samples, weighing up to 40 mg, were heated from 20 to
1100°C under nitrogen gas flow at a heating rate of 10°C/min. All TGA/DTG curves,
Chapter 3 80
mass changes, and temperature peaks were calculated using a TGA software. The amount
of BW and degree of hydration were related based on the Powers’ relation given in Eq.
3-1 (Loukili et al., 1999, Mounanga et al., 2004):
100)()((%) ×=
∞tWtBWα Eq. 3-1
Where α is the degree of hydration, is the amount of chemically bound water at
time t and is the amount of water required for complete hydration. For the binder
used in this study, was evaluated to be approximately 0.223 (g/g binder) using 18
months age specimens. The starting temperature for calculating the BW ranges from 103
to 112°C according to DTG results, which was within the temperature range used in
previous studies on UHPC (Loukili et al., 1999). However, if the tested sample is dried
well, then the induced error by the difference in the starting temperature will be
negligible (Pane and Hansen, 2005). Furthermore, the measured was corrected to
consider the mass loss due to CO
)(tBW
)( ∞tW
)( ∞tW
)(tBW
2 release between 600 and 780°C as a result of calcite
decomposition (Mounanga et al., 2004, Pane and Hansen, 2005).
3.4.3.2 Strain Measurements
For each mixture, four (25×25×280 mm) specimens per curing condition were made
according to ASTM C 157 (Standard Test Method for Length Change of Hardened
Hydraulic-Cement Mortar and Concrete). Immediately after demolding, specimens for
autogenous strain measurements were wrapped with four layers of polyethylene sheets
and a layer of paraffin wax membrane to prevent moisture loss. Specimens for total strain
measurements were exposed to different curing conditions inside the walk-in
environmental chamber. The measured unrestrained one-dimensional deformations have
Chapter 3 81
been measured using a comparator provided by a dial gauge with an accuracy of 10
µm/m. Moreover, Type-T thermocouples were inserted in the specimens to monitor
internal temperature changes from the onset of tests. Small cross-section prismatic
(25×25×280 mm) specimens were chosen to reduce the moisture gradients effect induced
by drying (Lam, 2005) and to assure quick dissipation of the hydration heat (Bao-guo, et
al., 2007, Baroghel-Bouny et al., 2006). In addition, the high surface area to volume ratio
of the tested specimens was intended to enhance the effect of drying conditions through
reaching moisture equilibrium between the specimen and it’s surrounding in a shorter
time.
3.4.3.3 Moisture Loss
Prismatic specimens (25×25×280 mm) were made for each mixture and demolded at the
time of starting total strain measurements. Prisms were transferred to the walk-in
environmental chamber after measuring the initial mass of each prism using a balance
with an accuracy of 0.01 g. The mass measurements were taken for all prisms along with
measurements of the total strains. Each mass loss test result in this study represents the
average value obtained on four identical prisms.
3.4.3.4 Coefficient of Thermal Expansion
The coefficient of thermal expansion (CTE) was measured based on the temperature
variations method introduced by (Cusson and Hoogeveen, 2006, Cusson and Hoogeveen,
2007). In this method, tested specimens were subjected to a realistic temperature history
including small range temperature cycles. Three sealed concrete prisms (25×25×280 mm)
from each mixture were tested simultaneously in the walk-in environmental chamber
from the demolding time up to 48 hrs during which an important variation in CTE
Chapter 3 82
occurred before it stabilized thereafter. During the 25-30°C temperature cycles, the
temperature was maintained constant at each temperature step for 3 hrs, including a 15
minute ramp between each temperature step, thus resulting in 4 completed cycles per day.
Each specimen was fixed to an invar frame at one end and was free to deform
over two invar rods at the other end. The longitudinal deformation of the concrete prism
was measured using a high accuracy (0.001mm) LVDT located at the ends of the invar
frame. A foam rubber pad was placed between the test setup and the floor of the walk-in
environmental chamber to eliminate ambient vibrations. The concrete temperature was
measured using Type-T thermocouples embedded at the centre of the UHPC specimens.
In addition, the temperature inside the walk-in environmental chamber was monitored
using other thermocouples. Figure 3-1 presents a diagram of the test setup. The test setup
was calibrated under similar experimental conditions in order to eliminate undesirable
temperature effects induced by LVDT sensors and the metal and geometry of the test
apparatus. All measurements were corrected based on the calibration factor obtained (Eq.
3-2). Further details on the CTE calculation method can be found in (Cusson and
Hoogeveen, 2006).
T610mc ∆−×α+κε=ε Eq. 3-2
Where cε is the corrected strains, mε is the measured strains, κ is a calibration factor
(0.9925), and α is the coefficient of thermal expansion of the calibrated test apparatus
(1.267 x10 -6 /°C).
Chapter 3 83
Figure 3-1: Coefficient of thermal expansion test setup.
3.5. RESULTS AND DISCUSSION
The discussion below will focus on the influence of drying conditions, including the
ambient temperature and relative humidity, on the measured autogenous strains of UHPC
specimens. The thermal strain evaluated based on the temperature change at the centre of
the tested specimens and the corresponding CTE was excluded from the total measured
strain. (Note: CTE results are included in Appendix A).
3.5.1. Compressive Strength
Compressive strength is considered as a key property of UHPC. Results indicate that
curing conditions, including the availability of moisture and temperature profile, can
significantly affect the early-age compressive strength of UHPC. Figure 3-2 shows the
effect of the curing temperature on the compressive strength development of UHPC.
Higher curing temperature resulted in higher compressive strength, especially during the
first 24 hrs. The compressive strength at 24 hrs for specimens cured at 10 and 20°C
achieved only 45% and 70% of that cured at 40°C, respectively. However, the rate of
strength gain was lower at later age for specimens cured at 40°C compared with that at 10
Chapter 3 84
and 20°C. For instance, the strength gain at 40°C between ages 3 and 7 days was only
about 9%, while it was about 29% at 20°C.
0
20
40
60
80
100
120
140
10°C-60%RH 20°C-60%RH 40°C-60%RHAmbient condition
Com
pres
sive
str
engt
h (M
Pa) 1 day
3 days7 days
Figure 3-2: Effect of curing temperature on compressive strength for mixtures with
w/c=0.25.
During early-age, higher ambient humidity led to higher compressive strength,
while at latter age, this effect started to diminish as shown in Fig. 3-3. The strength
gained during the first 24 hrs at 20°C and 40% ambient humidity was about 25% lower
than that gained under 80% ambient humidity, while this difference was only 10 and 7%
at 3 and 7 days, respectively. This is likely due to the fact that for UHPC a lower amount
of hydration is required to gain higher strength since space between cement particles and
porosity to be filled by hydration products are relatively low. In addition, at later ages the
depercolation of capillary pores will reduce moisture exchange between the core of the
specimens and it’s surrounding (Bentz and Garboczi, 1991).
Chapter 3 85
0
20
40
60
80
100
20°C-40%RH 20°C-60%RH 20°C-80%RHAmbient condition
Com
pres
sive
str
engt
h (M
Pa) 1 day
3 days7 days
Figure 3-3: Effect of ambient humidity on compressive strength for mixtures with
w/c=0.25.
Generally, decreasing the w/c resulted in greater compressive strength. However,
the compressive strength results for the control mixture with w/c=0.22 were comparable
to that of the control mixture with w/c=0.25 as shown in Fig. 3-4. For instance, the
percentage difference in early-age compressive strength between control mixtures with
w/c=0.22 and 0.25 after 7 days under different curing conditions, ranged from about (-
5.0%) to (+4.0%). Moreover, this difference decreased further at 28 days (ranged from (-
3.2%) to (+2.0%)). Not achieving significantly greater compressive strength with lower
w/c can be ascribed to the agglomeration of SF particles which reduced its effectiveness.
The agglomerated grains of silica fume have less effective pozzolanic activity than that of
individual grains (Yajun and Cahyadi, 2003). In addition, the size of such agglomerates
often exceeds that of Portland cement particles, thus limiting the benefits attributed to the
fine particle filler effect (Diamond and Sahu, 2006), leading to higher porosity. However,
Chapter 3 86
such porosity may provide more space for further hydration products, leading to retrieval
of strength at later ages. It was observed by (Korpa and Trettin, 2004, Diamond et al.,
2004) that at later age, the pozzolanic reaction of large silica fume particles can attain a
considerable degree, contributing remarkably to the development of mechanical
properties.
020406080
100
120140160180200
10°C-60%RH
10°C-80%RH
20°C-40%RH
20°C-60%RH
20°C-80%RH
40°C-40%RH
40°C-60%RH
40°C-80%RH
Curing conditions
Com
pres
sive
str
engh
t (M
Pa) w/c=0.25- 7Days
w/c=0.22- 7Daysw/c=0.25- 28Daysw/c=0.22- 28Days
Figure 3-4: Effect of w/c on compressive strength.
3.5.2. Total Strain
3.5.2.1 Influence of Curing Temperature and Ambient Humidity on Total Strain
The total amount of strain mainly depends on the exposure conditions and concrete
mixture design (Mönnig and Lura, 2007). Figure 3-5 illustrates the effect of exposure
conditions on the measured total strains for control UHPC mixtures. At higher
temperature, specimens were found to exhibit higher total strains compared to that of
specimens exposed to lower temperature, regardless of the RH. In addition, under the
Chapter 3 87
same exposure temperature, lowering the RH increased the total strains. This increase
was more significant at high temperature as shown in Fig. 3-5. Moreover, mixtures with
w/c=0.22 showed lower total strain compared to that of mixtures with w/c=0.25. This can
be attributed to lower water content of w/c=0.22 mixtures and consequently lower
evaporable water.
3.5.3. Autogenous Strain
3.5.3.1 Effect of Curing Temperature and w/c
Figure 3-6 shows the autogenous strain for the control mixture (w/c=0.25) under sealed
conditions for the first 7 days at different curing temperatures. Increasing the temperature
from 20 to 40°C accelerated the autogenous strain rate during the first 24 hours, which
could be explained by the acceleration of hydration reactions, resulting in higher
chemical shrinkage. The latter is considered as the main driving force for shrinkage until
an internal rigid skeleton is formed (Jensen and Hansen, 1999, Mounanga et al., 2004,
Acker, 2004, Kamen et al., 2008).
While the rate of autogenous strain at 40°C started to decrease after the first day
of hydration, this took more than 4 days at 20°C. This could be attributed to the dual
action of temperature on the microstructure of the cementitious matrix. High curing
temperature motivates the development of a strong solid skeleton (Kamen et al., 2008),
which resists autogenous strain (Yang et al., 2005, Esping, 2007). Simultaneously, the
high curing temperature results in a coarser porous structure (larger pore radius) (Neville
and Brooks, 1991, Cao and Detwiler, 1995), leading to lower capillary stresses as shown
by the Kelvin-Laplace equations (Hua et al., 1995). Consequently, autogenous strain will
develop with a lower rate (Kamen et al., 2008).
Chapter 3 88
-1200
-1000
-800
-600
-400
-200
0
Tota
l Str
ain
(µε)
10°C20°C40°C
RH=40% RH=80%RH=60%Ambient humidity(a)
-1200
-1000
-800
-600
-400
-200
0
Tota
l Str
ain
(µε)
10°C20°C40°C
Ambient humidityRH=40% RH=80%RH=60%
(b)
Figure 3-5: Total strains Vs. ambient humidity at different curing temperatures for
a) w/c=0.25 and b) w/c=0.22 mixtures at age of 7-days.
Chapter 3 89
-600
-500
-400
-300
-200
-100
00 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
40°C
10°C
20°C
Figure 3-6: Autogenous strains for the control mixture (w/c=0.25) at 10, 20 and
40°C.
At 10°C, the developed strain showed a near plateau after a few hours from the
setting time, then started to expand slightly, before autogenous shrinkage strain increased
again. This expansion could not be attributed to the thermal effect of hydration reactions
since the specimen cross-section was small (Baroghel-Bouny et al., 2006), but could
occur likely due to the formation of ettringite (Loukili et al., 1999, Baroghel-Bouny et
al., 2006, Bentz et al., 2001a).
Generally, decreasing the w/c results in greater autogenous strains (Holt, 2001,
Zhang et al., 2003, Holt and Leivo, 2004). However, comparable strains were observed
for mixtures with w/c=0.22 and 0.25 at different temperatures as shown in Fig. 3-7. This
can be ascribed to the high SF content (>25%), which was reported to reduce the
autogenous strains for extremely low w/b mixtures as a result of the inadequate packing
Chapter 3 90
and agglomeration of SF particles, which consequently leads to higher porosity
(Baroghel-Bouny and Kheirbek, 2000, Yajun and Cahyadi, 2003).
-700
-600
-500
-400
-300
-200
-100
0
1000 20 40 60 80 100 120 140 160 18
Age (hours)A
utog
enou
s St
rain
(µε)
0
w/c=0.22- 10°Cw/c=0.22- 20°Cw/c=0.22- 40°Cw/c=0.25- 10°Cw/c=0.25- 20°Cw/c=0.25- 40°C
Figure 3-7: Autogenous strains for mixtures with w/c=0.22 and 0.25 at different
temperatures.
Furthermore, for high SF/cement mixtures, calcium hydroxide (CH) crystals
dissolve to provide calcium ions during the pozzolanic reaction. If the resulting calcium
silicate hydrate (CSH) products do not fill the space of dissolved CH crystals, porosity
can be created (Baroghel-Bouny and Kheirbek, 2000). This can partially explain the
absence of expansion at 10°C for w/c=0.22 mixtures since more pore space could have
been created. Figure 3-8 shows SEM and EDX microanalysis results (conducted at
Surface Science Western at The University of Western Ontario), which substantiate this
hypothesis in w/c=0.22 UHPC control mixtures. Spatial elemental distribution by EDX
revealed high intensity of silicon (Si) in the circular body representing an agglomerated
Chapter 3 91
SF particle. The agglomerated SF particle is encapsulated by a rich media of calcium
(Ca), silicon (Si) and oxygen (O) forming a halo of CSH.
(a)
(b)
Figure 3-8: Scanning electron micrograph for w/c =0.22 UHPC specimens: (a)
agglomerated silica fume encapsulated with C-S-H, (b) elemental distribution.
3.5.4. Degree of Hydration
3.5.4.1 Under Sealed Condition
The hydration of cement is highly sensitive to the w/c and curing temperature (Mounanga
et al., 2004, Lothenbach et al., 2007, Escalante-Garcia and Sharp, 2000, Slamecka and
Škvara, 2002). The hydration kinetics increases with increasing w/c and/or curing
temperature (Mounanga et al., 2004). TGA was used to investigate the hydration
evolution for samples taken from the cementitious matrix of the tested UHPC specimens.
The measured degree of hydration versus age is shown in Fig. 3-9. For all mixtures, the
Chapter 3 92
hydration rate increased with increasing temperature, which is in agreement with
previous work (Mounanga et al., 2004, Escalante-Garcia and Sharp, 2000). Reducing the
w/c from 0.25 to 0.22 did not cause a significant reduction in the degree of hydration
(Fig. 3-9(b)). This can be attributed to the fact that at very low w/c, the content of un-
hydrated phases, which could exceed 42% (Acker, 2004), may restrict the further
development of hydration products (Lothenbach et al., 2007).
10
20
30
40
50
60
70
0 20 40 60 80 100 120 140 160 180Age (hours)
Hyd
ratio
n D
egre
e (%
)
10°C20°C40°C
w/c: 0.25
(a)
Figure 3-9: Evaluation of degree of hydration Vs. age for control mixtures a)
w/c=0.25 and b) w/c= 0.22.
Chapter 3 93
10
20
30
40
50
60
70
0 20 40 60 80 100 120 140 160 180Age (hours)
Hyd
ratio
n D
egre
e (%
)
10°C20°C40°C
w/c:0.22
(b)
Figure 3-9 Contd’: Evaluation of degree of hydration Vs. age for control mixtures a)
w/c=0.25 and b) w/c= 0.22.
3.5.4.2 Under Drying Condition
Concrete is usually exposed to ambient air at early-ages, and a moisture exchange
between concrete and its environment occurs. Hence, the amount of water available for
hydration reactions will be altered. Generally, a reduction in water content at drying
conditions takes place, with an adverse effect on the hydration process of cement.
Accordingly, the amount of hydration products and BW will be changed (Bentz et al.,
2001b). Figure 3-10 shows cross sections for sealed and exposed to drying specimens.
The difference in the colors would indicate the difference in the hydration degree
between the two specimens.
Chapter 3 94
a) b)
Figure 3-10: Cross sections for a) sealed and b) exposed to drying specimens after 7 days.
Figure 3- 11 (a,b,c) shows the effect of the curing conditions on the BW. At
10°C, the change in RH induced a slight difference in the degree of hydration achieved.
This could be attributed to the low rate of evaporation and hydration (Escalante-Garcia
and Sharp, 2000), which slowed down the reduction of internal RH (Jensen and Hansen,
1999). In addition, the rate of hydration seemed to be similar over the investigated period.
At higher curing temperatures, the RH had a pronounced effect on both the rate of
hydration and the degree of hydration achieved. Low RH enhances the moisture transfer
from the specimen to its surrounding, which disturbs the hydration process, leading to a
lower degree of hydration. Generally, curing at high temperature induces a competition
for the mixing water between evaporation and hydration. High curing temperatures
usually result in a more porous and continuous pore structure, which facilitates
evaporation (Cao and Detwiler, 1995). Conversely, high curing temperature enhances the
hydration rate causing more water to bond chemically and physically to hydration
products, leaving a lower amount of evaporable water (Mounanga et al., 2004). The rate
of change of the internal RH, which is controlled by the hydration rate (Jensen and
Chapter 3 95
Hansen, 1999) and external RH level, will dominate the water exchange between the
tested specimen and its surrounding.
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
10°C-60% RH10°C-80% RH10°C-Sealed
(a)
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
20°C-40% RH20°C-60% RH20°C-80% RH20°C-Sealed
(b)
Figure 3-11: Change in BW Vs. ambient humidity for w/c=0.25 control mixtures at
a) 10°C, b) 20°C, and c) 40°C.
Chapter 3 96
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
40°C-40% RH40°C-60% RH40°C-80% RH40°C-Sealed
(c)
Figure 3-11 Cont’d: Change in BW Vs. ambient humidity for w/c=0.25 control
mixtures at a) 10°C, b) 20°C, and c) 40°C.
BW results were consistent with mass loss measurements as shown in Fig. 3-12
(a,b,c). While UHPC specimens cured at 10°C showed a slight increase in mass loss with
decreasing RH, significant variations in mass loss were observed at higher temperature.
Increasing the curing temperature and decreasing the RH enhanced the evaporation rate,
resulting in higher mass loss, which is in agreement with previous studies (Almusallam,
2001, Bissonnette et al., 1999, Al-Saleh and Al-Zaid, 2006).
Chapter 3 97
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
10°C-60% RH10°C-80% RH
(a)
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
20°C-40% RH20°C-60% RH20°C-80% RH
(b)
Figure 3-12: Change in mass loss Vs. ambient humidity for w/c=0.25 control
mixtures at a) 10°C, b) 20°C, and c) 40°C.
Chapter 3 98
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
40°C-40% RH40°C-60% RH40°C-80% RH
(c)
Figure 3-12 Cont’d : Change in mass loss Vs. ambient humidity for w/c=0.25 control
mixtures at a) 10°C, b) 20°C, and c) 40°C.
3.5.5. Correlation between Autogenous Strain and Achieved Degree of Hydration
under Sealed Condition
Autogenous strains were plotted in Fig. 3-13 for mixtures with w/c=0.22 and 0.25 cured
at 10, 20 and 40°C as a function of the amount of BW determined by TGA tests. The
amount of BW was fixed at zero at the start of autogenous strain measurements. Thus,
BW will be considered as the relative chemically bound water (RBW) in the proposed
relationship. The experimental results revealed a good correlation between the measured
autogenous strains and the RWB, as shown in Fig. 3-13.
Chapter 3 99
R2 = 0.93
-800
-700
-600
-500
-400
-300
-200
-100
00 0.02 0.04 0.06 0.08 0.1
RBW (g/g cement)
Aut
ogen
ous
Stra
in (µ
ε)
w/c=0.22 - 10°Cw/c=0.22 - 20°Cw/c=0.22 - 40°Cw/c=0.25 - 10°Cw/c=0.25 - 20°Cw/c=0.25 - 40°C
Figure 3-13: Correlation between autogenous strains and RBW for control mixtures
cured at 10, 20 and 40°C.
3.5.6. Predicting Drying and Autogenous Strains under Different Curing
Conditions
To investigate the effect of drying conditions on the contribution of both drying and
autogenous strains to the total strains, the drying strains were determined based on the
superposition principle and the evaluated degree of hydration.
3.5.6.1 Superposition Principle
The superposition principle is based on the assumption that autogenous and drying strains
could be superimposed, and that autogenous strains in an exposed specimen are similar to
that in a sealed specimen. Thus, drying strains can be determined by subtracting the
autogenous strains in a sealed specimen from the total strains after applying the thermal
Chapter 3 100
strains correction (Tazawa et al., 2000). Figure 3-14 illustrates the drying strains in the
control specimens under different curing conditions. As the RH increased, drying strains
decreased, which is consistent with previous results (Bao-guo et al., 2007, Hansen, 2005,
Almusallam, 2001, Holt and Leivo, 2000, Tazawa and Miyazawa, 1999, Bazant and
Wittmann, 1982).
The effect of the temperature on drying strains at different RH did not exhibit a
clear trend. At RH=40%, drying strains increased as the curing temperature increased,
which is consistent with previous observation (Almusallam, 2001). At RH=60 and 80%,
specimens cured at 10°C showed higher drying strains than that of specimens cured at 20
and 40°C. These results are in agreement with measurements performed in a previous
study at 5, 20 and 30°C (Weiss and Berke,2003). It was argued that the delay in
developing a strong microstructure to resist deformations at low curing temperature was
the main reason for this behaviour.
-450
-350
-250
-150
-50
50
Dry
ing
Stra
in (µ
ε)
10°C20°C40°C
RH=40% RH=80%RH=60%Ambient humidity
Figure 3-14: Drying strains for w/c=0.25 control mixtures based on the
superposition principle at age of 7-days.
Chapter 3 101
At RH=80% and 20°C, the calculated drying strains based on the superposition
principle indicated swelling, which confirms that autogenous strains were higher than the
total strains. Similar results were reported by others (Tazawa et al., 2000, Tazawa and
Miyazawa, 1999) who attributed this behaviour to the moisture movement from the high
RH environment into the concrete specimens, driven by the lower intrinsic humidity
developed due to the effect of self-desiccation.
3.5.6.2 Degree of Hydration Method
In the degree of hydration method, autogenous strains under different exposure
conditions were evaluated based on the achieved degree of hydration (i.e. amount of
RBW) of the exposed specimen and the relationship between the degree of hydration and
the corresponding autogenous strain under sealed condition (Kovler and Zhutovsky,
2006). Consequently, the reductions in the autogenous strains due to the drying effect are
considered. Figure 3-15 shows the autogenous strains evaluated based on the degree of
hydration at different exposure conditions. It can be observed that autogenous strains
calculated based on the superposition principle under drying conditions were
overestimated as compared with that calculated using the degree of hydration. The
overestimation ratio (β ) calculated by Eq. 3-3 is listed in Table 3-4, in which
autogenous strains at 7 days are used for calculations.
w
ws
εεεβ −
= Eq. 3-3
Where sε is the autogenous strain under sealed conditions, wε is the autogenous strain
under drying conditions obtained based on the evaluated degree of hydration. The lower
Chapter 3 102
the RH, the more pronounced was the difference between the autogenous strains obtained
by the two methods. It should be noted that the early-age exposure to drying increased the
overestimation ratio, since it interrupts the hydration process and the development of
denser microstructure, thus resulting in higher porosity.
-1000
-800
-600
-400
-200
0
Aut
ogen
ous
Stra
in (µ
ε)
10°C20°C40°C
RH=40% RH=80%RH=60% SealedAmbient humidity
Figure 3-15: Autogenous strains evaluated based on achieved degree of hydration
for w/c=0.25 UHPC mixtures at age 7 days.
Table 3-4: Overestimation ratios for autogenous strains under different curing
conditions.
Overestimation ratio
Mixture Temp.°C/RH% 40 60 80
10 ‐‐‐‐* 0.78 0.46
20 1.26 0.98 0.45 w/c=0.25
40 1.74 1.17 0.63
10 ‐‐‐‐* 0.80 0.51
20 1.32 0.97 0.49 w/c=0.22
40 1.89 1.22 0.65
* Condition is not feasible.
Chapter 3 103
Figure 3-16 shows the drying strains determined by subtracting autogenous
strains from the total measured strains, after considering the reduction in the autogenous
strains due to the drying effect. Comparing Figs. 3-14 and 3-16, a significant increase in
the contribution of drying strains to the total measured strains can be observed. For
instance, at 40°C and RH=40%, the calculated drying strains based on the superposition
principle represented about 40% of the total measured strains, compared to about 78%
based on the evaluated degree of hydration. Moreover, no swelling was observed at high
RH, which is consistent with the expected denser structure for the tested mixtures
(Esping, 2007). Generally, a good agreement exists between the determined drying
strains and mass loss results (Fig. 3-12(a,b,c) and 3-16).
-1000
-800
-600
-400
-200
0
Dry
ing
Stra
in (µ
ε)
10°C20°C40°C
RH=40% RH=80%RH=60%Ambient humidity
Figure 3-16: Drying strains evaluated based on achieved degree of hydration for
w/c=0.25 UHPC mixtures at age 7 days.
Chapter 3 104
3.6. CONCLUSIONS
Using the amount of chemically bound water to assess autogenous and drying strains in
ultra high-performance concrete specimens was explored in this chapter. The following
conclusions can be drawn from the experimental results:
1) Ambient conditions, including temperature and relative humidity, can
significantly affect the development of early-age compressive strength of UHPC
cast in-situ, especially in cold temperature for which the achieved strength was
only about 52% of the target strength.
2) Autogenous and drying strains in UHPC specimens are dependent phenomena;
therefore the behaviour of autogenous strains under sealed conditions will differ
from that under drying conditions.
3) Autogenous shrinkage strain and drying shrinkage strain of concrete under drying
conditions can be separated from total shrinkage strain based on the evaluated
degree of hydration based on BW
4) Applying the superposition principle without considering the effect of drying
conditions will result in overestimating autogenous strains, especially in arid
conditions.
5) For thin UHPC sections, drying dominates the total deformation and reduces the
development of autogenous strains, leading to lower autogenous contribution to
the total deformation.
Chapter 3 105
3.7. REFERENCES
Acker, P., (2004), “Swelling, shrinkage and creep: a mechanical approach to cement hydration,” Materials and Structures, Vol. 37, No. 4, pp. 237-243.
Aïtcin, P.C., (1999), “Demystifying autogenous shrinkage,” Concrete International, Vol.
21, No. 11, pp. 54-56. Al-Saleh, S.A.and Al-Zaid, R.Z., (2006), “Effects of drying conditions, admixtures and
specimen size on shrinkage strains,” Cement and Concrete Research, Vol. 36, No. 10, pp. 1985-1991.
Almusallam, A.A., (2001), “Effect of environmental conditions on the properties of fresh
and hardened concrete,” Cement and Concrete Composites, Vol. 23, No. 4-5, pp. 353-361.
Bao-guo, M., Xiao-dong, W., Ming-yuan, W., Jia-jia, Y. and Gao, X-j., (2007), “Drying
Shrinkage of Cement-Based Materials Under Conditions of Constant Temperature and Varying Humidity,” Journal of China University of Mining and Technology, Vol. 17, No. 3, pp. 428-431.
Baroghel-Bouny, V. and Kheirbek, A., (2000), “Effect of mix-parameters on autogenous
deformations of cement pastes - Microstructural interpretations,” In: V. Baroghel-Bouny and P.C. Aïtcin, Editors, Proceedings of the Int. RILEM Workshop on Shrinkage of Concrete “Shrinkage 2000”, Paris, France, RILEM Publication.
Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafai, N., (2006),
“Autogenous deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp. 123-136.
Bazant, Z.P. and Wittmann, F.H, (1982). Creep and Shrinkage in Concrete Structures,
John Wiley and Sons, Chichester. Bentz, D.P. and Garboczi, E.J., (1991), “Percolation of phases in a three-dimensional
cement paste microstructural model,” Cement and Concrete Research, Vol. 21, No. 2-3, pp. 325-344.
Bentz, D.P., Hansen, K.K., Madsen, H.D., Vallee, F. and Griese, E.J., (2001b),
“Drying/hydration in cement pastes during curing,” Materials and Structures, Vol. 34, No. 9, pp. 557-565.
Bentz, D.P., Jensen, O.M., Hansen, K.K., Olesen, J.F., Stang, H. and Haecker, C-J.,
(2001a), “Influence of cement particle-size distribution on early age autogenous strains and stresses in cement-based materials,” Journal of American Ceramic Society, Vol. 84, No. 1, pp.129-135.
Chapter 3 106
Bissonnette, B., Pierre, P. and Pigeon, M., (1999), “Influence of key parameters on drying shrinkage of cementitious materials,” Cement and Concrete Research, Vol. 29, No. 10, pp.1655-1662.
Cao, Y. and Detwiler, R.J., (1995), “Backscattered electron imaging of cement pastes
cured at elevated temperatures,” Cement and Concrete Research, Vol. 25, No. 3, pp. 627-638.
Cheyrezy, M., Maret, V. and Frouin, L., (1995), “Microstructural analysis of RPC
(Reactive Powder Concrete),” Cement and Concrete Research, Vol. 25, No. 7, pp. 1491-1500.
Cusson, D. and Hoogeveen, T.J., (2006), “Measuring early-age coefficient of thermal
expansion in high-performance concrete,” In: Proceeding of International RILEM Conference on Volume Changes of Hardening Concrete: Testing and Mitigation, Lyngby, Denmark, RILEM Publication.
Cusson, D. and Hoogeveen, T.J., (2007), “An experimental approach for the analysis of
early-age behaviour of high-performance concrete structures under restrained shrinkage,” Cement and Concrete Research, Vol. 37, No. 2, pp. 200-209.
Diamond, S., Sahu, S. and Thaulow, N., (2004), “Reaction products of densified silica
fume agglomerates in concrete,” Cement and Concrete Research, Vol. 34, No. 9, pp.1625-1632.
Diamond, S. and Sahu, S., (2006), “Densified silica fume: particle sizes and dispersion in
concrete,” Materials and Structures, Vol. 39, No. 9, pp. 849-859. Escalante-Garcia, J.I. and Sharp, J.H., (2000), “The effect of temperature on the early
hydration of Portland cement and blended cements,” Advanced Cement Research., Vol. 12, No. 3, pp. 121-130.
Esping, O., (2007). “Early-age properties of self-compacting concrete - Effects of fine
aggregate and limestone filler,” PhD Thesis, Department of Civil and Environmental Engineering, Chalmers University of Technology.
Graybeal, B.A. and Hartmann, J.L., (2005), “Construction of an optimized UHPC vehicle
bridge,” Proceedings of the Seventh International Symposium on Utilization of high-strength/ high performance concrete, Washington D.C., USA, Vol. 2, pp.1109-1118.
Hansen, W., (2005), “Drying shrinkage mechanisms in portland cement paste,” Journal
of American Ceramic Society, Vol. 70, No. 50, pp. 323-328. Holt, E. and Leivo, M.T., (2000), “Methods of reducing early-age shrinkage,”
Proceedings of the International RILEM Workshop on Shrinkage of Concrete
Chapter 3 107
“Shrinkage 2000”, V. Baroghel-Bouny and P.C. Aïtcin (ed.), Paris, France, RILEM Publication, 435-448.
Holt, E., (2001) “Early Age Autogenous Shrinkage of Concrete,” PhD Thesis, University
of Washington. Holt, E. and Leivo, M., (2004), “Cracking risks associated with early age shrinkage,”
Cement and Concrete Composites, Vol. 26, No. 5, pp. 521-530. Holt, E., (2005), “Contribution of mixture design to chemical and autogenous shrinkage
of concrete at early ages,” Cement and Concrete Research, Vol. 35, No. 3, pp. 464-472.
Hua, C., Acker, P. and Ehrlacher, A., (1995), “Analyses and models of the autogenous
shrinkage of hardening cement paste: I. Modelling at macroscopic scale,” Cement and Concrete Research, Vol. 25, No. 7, pp. 1457-1468.
Ichinomiya, T., Hishiki, Y., Ohno, T., Morita, Y. and Takada, K., (2005), “Experimental
study on mechanical properties of ultra-high-strength concrete with low-autogenous-shrinkage,” American Concrete Institution, Special Publication, No. SP-228, pp. 1341-1352.
Jensen, O.M. and Hansen, P.F., (1996), Autogenous deformation and change of the
relative humidity in silica fume-modified cement paste, ACI Materials Journal, Vol. 93, No. 6, pp.1-5.
Jensen, O.M. and Hansen, P.F., (1999), “Influence of temperature on autogenous
deformation and relative humidity change in hardening cement paste,” Cement and Concrete Research, Vol. 29, No. 4, pp. 567-575.
Kamen, A., Denarié, E., Sadouki, H. and Brühwiler, E., (2008), “Thermo-mechanical
response of UHPFRC at early age - Experimental study and numerical simulation,” Cement and Concrete Research, Vol. 38, No. 6, pp. 822-831.
Katrin, H., Marco, V., Emmanuel, D. and Eugen, B., (2006), “Development of the
mechanical properties of an Ultra-High Performance Fiber Reinforced Concrete (UHPFRC),” Cement and Concrete Research, Vol. 36, No.7, pp. 1362-1370.
Kjellsen, K.O. and Detwiler, R.J., (1992), “Reaction kinetics of Portland cement mortars
hydrated at different temperatures,” Cement and Concrete Research, Vol. 22, No. 1, pp. 112-120.
Korpa, A. and Trettin, R., (2004), “The use of synthetic colloidal silica dispersions for
making HPC and UHPC systems, preliminary comparison results between colloidal silica dispersions and silica fumes (SF),” In: Proceedings of the Int. Symposium on Ultra High Performance Concrete, Kassel, Germany, pp. 155-163.
Chapter 3 108
Kovler, K. and Zhutovsky, S., (2006), “Overview and future trends of shrinkage research,” Materials and Structures, Vol. 39, No. 9, pp. 827-847.
Lam, H., (2005), “Effects of internal curing methods on restrained shrinkage and
permeability,” Master Thesis., University of Toronto. Lam, L., Wong, Y.L. and Poon, C.S., (2000), “Degree of hydration and gel/space ratio of
high-volume fly ash/cement systems,” Cement and Concrete Research, Vol. 30, No. 5, pp. 747-756.
Lothenbach, B., Winnefeld, F., Alder, C., Wieland, E. and Lunk, P., (2007), “Effect of
temperature on the pore solution, microstructure and hydration products of Portland cement pastes,” Cement and Concrete Research, Vol. 37, No. 4, pp 483-491.
Loukili, A., Khelidj, A. and Richard, P., (1999), “Hydration kinetics, change of relative
humidity, and autogenous shrinkage of ultra-high-strength concrete,” Cement and Concrete Research, Vol. 29, No. 4, pp. 577-584.
Lura, P. and Jensen, O.M., (2007), “Measuring techniques for autogenous strain of
cement paste,” Materials and Structures, Vol. 40, No. 4, pp. 431-440. Mönnig, S. and Lura, P., (2007), “Superabsorbent Polymers - An Additive to Increase the
Freeze-Thaw Resistance of High Strength Concrete,” In: Advances in Construction Materials, Part V, Springer Berlin Heidelberg, pp. 351-358.
Mounanga, P., Khelidj, A., Loukili, A. and Baroghel-Bouny, V., (2004), “Predicting
Ca(OH)2 content and chemical shrinkage of hydrating cement pastes using analytical approach,” Cement and .Concrete Research, Vol. 34, No. 2, pp. 255-265.
Neville, A.M. and Brooks, J., (1991). Concrete Technology. Longman Scientific and
Technical, Harlow, Essex, UK. Pane, I. and Hansen, W., (2005), “Investigation of blended cement hydration by
isothermal calorimetry and thermal analysis,” Cement and Concrete Research, Vol. 35, No. 6, pp. 1155-1164.
Schachinger, I., Schubert, J. and Mazanec, O., (2004), “Effect of mixing and placement
methods on fresh and hardened ultra high performance concrete (UHPC),” Proceedings of the International Symposium on Ultra High Performance Concrete, Kassel, Germany, pp. 575-586.
Slamecka, T. and Škvara, F., (2002), “The effect of water to cement ratio on the
microstructure and composition of the hydration products of Portland cement pastes,” Ceramics − Silikáty, Vol. 46, No. 4, pp. 152-158.
Chapter 3 109
Tang, M.C., (2004), “High Performance Concrete – Past, Present and Future,” Proceedings of the International Symposium on Ultra High Performance Concrete, Kassel, Germany, pp. 3-9.
Tazawa, E., (1999). Autogenous Shrinkage of Concrete. Proceedings of the International
Workshop Organized by the Japan Concrete Institute, Hiroshima, Japan. Tazawa, E. and Miyazawa, S., (1999), “Effect of constituents and curing condition on
autogenous shrinkage of concrete,” In: Autogenous Shrinkage of Concrete: Proceedings of the International Workshop, Eiichi Tazawa, (ed), Taylor & Francis, pp. 269-280.
Tazawa, E., Sato, R., Sakai, E. and Miyazawa, S., (2000). Work of JCI committee on
autogenous shrinkage. In: Proceedings of the International RILEM Workshop on Shrinkage of Concrete “Shrinkage 2000”, V. Baroghel-Bouny and P.C. Aïtcin (ed.), Paris, France, RILEM Publ., pp. 21-40.
Weiss, J. and Berke, N. (2003), “Admixtures for reduction of shrinkage and cracking,”
In: Early Age Cracking in Cementitious Systems: Report of RILEM Technical Committee 181-EAS, Bentur A; (ed.), pp. 323–335.
Yang, Y., Sato, R. and Kawai, K., (2005), “Autogenous shrinkage of high-strength
concrete containing silica fume under drying at early ages,” Cement and Concrete Research, Vol. 35, No. 3, pp. 449-456.
Yajun, J. and Cahyadi, J.H., (2003), “Effects of densified silica fume on microstructure
and compressive strength of blended cement pastes,” Cement and Concrete Research, Vol. 33, No. 10, pp. 1543-1548.
Yuasa, N., Kasai, Y. and Matsui, I., (1999), “Inhomogeneous distribution of the moisture
content and porosity in concrete,” Proceedings of the International Conference, R.K. Dhir and M.J. McCarthy (Ed.), The University of Dundee, Scotland, United Kingdom, Thomas Telford Publication, pp. 93-101.
Youhua, Y., (2000), “Manufacturing Reactive Powder Concrete using common new
Zealand materials,” Master thesis, University of Auckland. Zhang, M.H., Tam, C.T. and Leow, M.P., (2003), “Effect of water-to-cementitious
materials ratio and silica fume on the autogenous shrinkage of concrete,” Cement and Concrete Research, Vol. 33, No. 10, pp.1687-1694.
Chapter 4 110
CHAPTER FOUR
INFLUENCE OF SHRINKAGE MITIGATION TECHNIQUES ON EARLY-AGE SHRINKAGE
INTERRELATION UNDER DRYING CONDITIONS*
In the preceding chapter (3), the combined effect of drying and autogenous shrinkage,
and their interaction mechanisms was investigated. It was proven that autogenous and
drying shrinkage are dependent phenomena. Reduction in the water content under drying
conditions, due to evaporation, had an adverse effect on hydration process and
development of autogenous shrinkage. However, the role of commonly used shrinkage
mitigation methods, including shrinkage-reducing admixture (SRA) and a superabsorbent
polymer (SAP), on these previous findings in Chapter 3 is not clear. Hence, this Chapter
provides fundamental investigation for the effect of ambient conditions on these
shrinkage mitigation methods mechanisms, suitable conditions for each mitigation
method and how to optimize its benefit in reducing shrinkage strains.
4.1. INTRODUCTION
General literature pertinent to UHPC self-desiccation, autogenous and drying shrinkage
behaviour during early-ages has been given in Chapter 3 (Section 3.1). It had pointed out
that early-age shrinkage and high cracking potential can defeat the purpose of using
UHPC. Hence, several shrinkage mitigation strategies have been proposed to reduce
shrinkage development and avoid crack formation. Among these, the most commonly
used is the addition of shrinkage-reducing admixtures (SRAs). SRAs directly influence
*A version of this chapter was published online in Materials and Structures. Some parts of this chapter were also published in the Eighth International Conference on Short and Medium Span Bridges, Niagara Falls, Ontario, Canada
Chapter 4 111
shrinkage by decreasing the surface tension of the pore solution, leading to lower
capillary stresses and consequently reduced shrinkage. Moreover, a new method using
superabsorbent polymers (SAPs) particles as a concrete admixture has been proposed.
SAPs are a group of polymeric materials that have the ability to absorb a significant
amount of liquid from the surrounding environment and to retain it within their structure
without dissolving. The SAP shrinkage mitigation technique is similar to using saturated
lightweight aggregate (LWA) particles but in a better controlled manner. However, the
efficacy of these mitigation techniques under field-like conditions is not guaranteed, since
it was evaluated under controlled laboratory conditions. In concrete exposed to field
conditions during early-ages, multi-mechanisms affect shrinkage behaviour
simultaneously (Yuasa et al., 1999). Hence, in this study a more realistic behaviour and
efficiency of SRA and SAP under simulated field-like conditions was investigated.
4.2. RESEARCH SIGNIFICANCE
Several studies have evaluated efficiency of various shrinkage mitigations techniques in
reducing shrinkage strains in UHPC under controlled curing condition without
accounting for the effect of the surrounding environment. A more realistic behaviour and
efficiency for SRA and SAP was evaluated based on degree of hydration approach
(described in previous chapter). Efficiency of SRA and SAP was found to be highly
influenced by exposure conditions. Field conditions should given a greater attention
while choosing the shrinkage mitigation method.
4.3. METHODOLOGY
Basically, the same methodology discussed in Chapter 3 was followed (see section 3.3).
Chapter 4 112
4.4. EXPERIMENTAL PROGRAM
4.4.1. Materials and Mixture Proportions
The materials used in this chapter were similar to that used in Chapter 3 (refer to Section
3.4.1). Four UHPC mixture incorporating 2% SRA or 0.6% SAP with w/c= 0.22 and 0.25
were tested in order to investigate the effect of both shrinkage mitigation techniques. The
chemical and physical properties of the used binders have been given in Chapter 3 (refer
to Table 3-1). A commercially available SRA composed mainly of poly-oxyalkylene
alkyl ether was used in this study. A 2% SRA by mass of cement was added as partial
replacement for mixing water, in agreement with previous works (Bentz et al., 2001a,
Bentz, 2006, Loser and Leemann, 2009). SAP is a covalently cross-linked polyacrylic
acid, suspension polymerized, white spherical particles with a particle size in the range of
100 to 140 µm. SAP was added at a rate of 0.6% by weight of cement as this value had
shown a great influence in diminishing shrinkage and reducing the cracking potential in
previous research (Jensen and Hansen, 2002, Lam, 2005, Lura and Jensen, 2007).
Mixtures with SAP contain an additional entrained water to offset self-desiccation. The
w/c of mixture incorporating SAP was increased to account for the amount of water that
will be absorbed by SAP particles to offset self-desiccation (Jensen and Hansen, 2001).
4.4.2. Environmental Conditions
The environmental conditions applied in this chapter were similar to that used in Chapter
3 (refer to Section 3.4.2).
Chapter 4 113
4.4.3. Preparation of Test Specimens and Testing Procedures
Testing samples were prepared and tested following the same procedure discussed in
Chapter 3 (refer to Section 3.4.3).
4.5. RESULTS AND DISCUSSION
4.5.1. Compressive Strength
Addition of SRA reduced the rate of cement hydration and strength development in
concrete. This led to a delay in setting and a reduction in the early-age compressive
strength compared to that of control specimens with similar w/c as shown in Fig. 4-1.
Moreover, reducing the ambient humidity had a minor effect on the achieved strength for
SRA mixtures compared to that for control mixtures. This can be ascribed to the sharp
drying front induced by SRA at low ambient humidity, which led to lower water loss
(Bentz et al., 2001a, Bentz, 2006).
Mixtures incorporating SAP exhibited lower compressive strength compared to that
of other mixtures, as shown in Fig. 4-1. However, SAP incorporated higher water
content, which enhanced the hydration process; it resulted in higher total porosity. At low
ambient humidity, water evaporates faster leading to more porosity and lower
development of hydration.
Chapter 4 114
0
20
40
60
80
100
120
140
60%RH 80%RH 40%RH 60%RH 80%RH 40%RH 60%RH 80%RH
Ambient Humidity (%)
Com
pres
sive
str
engt
h (M
Pa)
0
5
10
15
20
Com
pres
sive
str
engt
h (k
ips)
Control SRA SAP
T=10°C T=40°CT=20°C
Figure 4-1: Effect of adding 2% SRA and 0.6%SAP on compressive strength after 7
days for mixtures with w/c=0.25.
As expected, SRA reduced the 28 days compressive strength. The reduction in the
strength ranged between -5.8% to -11.0%. This is in agreement with ranges reported in
previous study on low w/c concrete (Weiss and Berke, 2003). The reduction in the
achieved strength can be attributed to the retardation effect induced by SRA addition. On
the other hand, mixtures incorporating SAP exhibited higher compressive strength
reduction (ranged between -12 to -22%) compared to SRA mixtures. Addition of SAP
induced higher and continuous porosity that facilitates water exchange with the
surrounding environments. As a result, higher porosity is required to be filled by
hydration product to gain strength compared to that of control mixtures. This is
consistent with previous findings which indicate that porosity has a dominate effect on
the achieved strength (Odler and Rößler, 1985, Mikhail et al., 1977).
Chapter 4 115
4.5.2 Total Strain
Generally, SRA mixtures exhibited lower total strain compared to that of control
mixtures (Fig. 3-5(a) & Fig. 4-2(a)). The reduction in total strain increased at higher
curing temperature and/or lower RH. For instance, the reduction in the total strain was
about 325 µε and 267 µε at (40°C and RH=40%) and (20°C and RH=40%) than that of
control mixtures, respectively. Conversely, SAP mixtures showed higher total strain
compared to that of control mixtures (Fig. 3-5(a) & Fig. 4-2(b)). Moreover, a higher total
strain was exhibited with higher curing temperature and/or lower RH. For instance, the
increase in the total strain was about 109 µε and 71 µε at (40°Cand RH=40%) and (20°C
and RH=40%) than that of control mixtures, respectively.
-1200
-1000
-800
-600
-400
-200
0
Tota
l str
ain
(µε)
10°C20°C40°C
RH=40% RH=80%RH=60%Ambient humidity(a)
Figure 4-2: Total strains Vs. ambient humidity at different curing temperatures for
w/c=0.25 mixtures at age of 7-days: a) SRA and b) SAP.
Chapter 4 116
-1200
-1000
-800
-600
-400
-200
0
Tota
l str
ain
(µε)
10°C20°C40°C
RH=40% RH=80%RH=60%Ambient humidity(b)
Figure 4-2 Contd’: Total strains Vs. ambient humidity at different curing
temperatures for w/c=0.25 mixtures at age of 7-days: a) SRA and b) SAP.
4.5.3. Autogenous Strain
4.5.3.1 Effect of Curing Temperatures
The effect of adding 2% SRA and 0.6% SAP on autogenous strains at 10, 20, and 40°C
for UHPC mixtures made with w/c=0.25 is shown in Fig. 4-3. It can be observed that
SRA and SAP reduced autogenous strains effectively, though SRA achieved more
significant reduction. Moreover, the effect of SRA was more significant at higher
temperature (40°C) where the reduction in autogenous strains was about 55% compared
with 34% and 32% at 10 and 20°C, respectively. This is believed to be due to the increase
of SRA concentration in the pore solution as a result of the accelerated hydration
reactions at high temperature. Indeed, hydration reactions consume mixing water, while
Chapter 4 117
SRA remains or would be consumed at a slower rate compared to that of water
(Rajabipour et al., 2008), thus leading to higher SRA concentration.
-600
-500
-400
-300
-200
-100
0Control SRA SAP
Aut
ogen
ous
Stra
in (µ
ε)
10°C 20°C 40°C
55%
34%
32%
71%
24%
28%
Figure 4-3: Autogenous strains after 7 days for mixtures with w/c=0.25.
Conversely, SAP appeared to be more effective at 10°C. It reduced autogenous
strains by about 71% compared to that of the control mixture. The reason for this
behaviour lays in the early expansion as shown in Fig. 4-4(a). To capture this effect, the
autogenous shrinkage curves were initiated to zero at the end of expansion as shown in
Fig. 4-4(b). The reduction in autogenous contraction after the early expansion was about
31 and 38% for SRA and SAP mixtures, respectively. Different phenomena are believed
to induce this early expansion including ettringite formation, expansion due to the
absorption of the internal curing water from SAP by the cement gel (Jensen and Hansen,
2002), and the relatively low stiffness of the specimens at early-age which permits such
an expansion (Baroghel-Bouny et al., 2006, Cusson and Hoogeveen, 2007). In addition,
Chapter 4 118
another mechanism can be hypothesized as follows: SAP particles release part of its
entrained water as hydration progresses (Wang et al., 2009), the vapour pressure in the
capillary pores increases and menisci curvature changes to have a large radius (Kovler,
1996). This will release the capillary surface tension, leading to capillary distension and
consequently expansion.
Indeed, autogenous shrinkage and expansion occur simultaneously during early-
ages. Early expansion along with autogenous shrinkage reduction due to SRA and/or
SAP can result in lower net autogenous strains. This is in agreement with previous
findings (Baroghel-Bouny et al., 2006, Cusson and Hoogeveen, 2007, Kamen et al.,
2008, Weiss et al., 2008).
-250
-200
-150
-100
-50
0
50
0 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
ControlSRASAP
(a)
Figure 4-4: Autogenous strains at 10°C: a) observed, and b) after expansion for
mixtures with w/c=0.25.
Chapter 4 119
-250
-200
-150
-100
-50
0
50
0 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
ControlSRASAP
(b)
Figure 4-4 Contd’: Autogenous strains at 10°C: a) observed, and b) after expansion
for mixtures with w/c=0.25.
4.5.3.2 Effect of w/c under Different Curing Temperatures
It was reported that SRA is more effective in reducing shrinkage at lower w/c
(D’Ambrosia, 2002). Conversely, SRA mixtures with w/c=0.22 showed higher
autogenous strain than that with w/c=0.25. In SRA mixtures, the role of SF
agglomeration and consequent creation of additional pore space can be explained as
follows: during early periods, the created pore space allows more hydration product
formation (e.g. ettringite formation (Bentz et al., 2001b, Bentz and Peltz, 2008)) without
causing high early expansion (see Fig. 4-5). Hence, SRA mixtures with w/c=0.22 showed
lower expansion compared to that of SRA mixtures with w/c=0.25, and thus higher
autogenous strain. At later ages, the effect of SRA on the surface tension of the pore
solution started to be the major mechanism controlling autogenous strain development.
Chapter 4 120
Hence, similar autogenous shrinkage strain rate was exhibited by both w/c=0.22 and 0.25
mixtures as shown in Fig. 4-6(a). Therefore, SF agglomeration can adversely affect SRA
mixtures during early-age. SAP mixtures exhibited similar trend to that of the control
mixtures as shown in Fig. 4-6(b). However, it is difficult at that point to evaluate the
effect of SF agglomeration on autogenous strain with respect to that of SAP. Further
research is needed to capture the effect of SAP mixtures without SF at similar level of
w/c.
Agglomerated SF particle
Figure 4-5: Scanning electron micrograph for agglomerated silica fume in SRA
mixture with w/c=0.22.
Chapter 4 121
-700
-600
-500
-400
-300
-200
-100
0
1000 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
w/c=0.22- 10°Cw/c=0.22- 20°Cw/c=0.22- 40°Cw/c=0.25- 10°Cw/c=0.25- 20°Cw/c=0.25- 40°C
(a)
-700
-600
-500
-400
-300
-200
-100
0
1000 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
w/c=0.22- 10°Cw/c=0.22- 20°Cw/c=0.22- 40°Cw/c=0.25- 10°Cw/c=0.25- 20°Cw/c=0.25- 40°C
(b)
Figure 4-6: Autogenous strains for mixtures with w/c=0.22 and 0.25 at different
temperatures a) SRA and b) SAP mixtures.
Chapter 4 122
4.5.4. Degree of Hydration
4.5.4.1 Under Sealed Condition
SRA retarded the hydration reactions at 10 and 40°C, as shown in Fig. 4-7(a). A slight
retardation in hydration reactions was observed at 20°C, indicating that the addition of
SRA did not affect the degree of hydration of specimens cured under a sealed condition.
This is consistent with previous results (Bentz et al., 2001a, Bentz, 2006)
SAP mixtures had a relatively higher water content (about 18% of w/c), thus
achieving a higher degree of hydration (Jensen and Hansen, 2002, Jensen and Hansen,
2001), as shown in Fig. 4-7(b). The increase in hydration rate was more pronounced at
40°C, which can be attributed to the thermo-activated characteristics of the hydration
process (Mounanga et al., 2004).
10
20
30
40
50
60
70
0 20 40 60 80 100 120 140 160 180Age (hours)
Hyd
ratio
n D
egre
e (%
)
10°C20°C40°C
w/c:0.25 + 2%SRA
(a)
Figure 4-7: Evaluation of degree of hydration Vs. age for a) w/c=0.25 + 2% SRA,
and b) w/c=0.25 + 0.6% SAP UHPC mixtures cured at 10, 20 and 40°C.
Chapter 4 123
10
20
30
40
50
60
70
0 20 40 60 80 100 120 140 160 180Age (hours)
Hyd
ratio
n D
egre
e (%
)
10°C20°C40°C
w/c:0.25 + 0.6%SAP
(b)
Figure 4-7 Contd’: Evaluation of degree of hydration Vs. age for a) w/c=0.25 + 2%
SRA, and b) w/c=0.25 + 0.6% SAP UHPC mixtures cured at 10, 20 and 40°C.
4.5.4.2 Under Drying Condition
4.5.4.2.1 SRA mixtures
Figure 4-8 (a,b,c) shows the effect of curing conditions on the BW for mixtures
incorporating 2% SRA. At 20°C, specimens cured at RH=40% showed higher BW
compared to that of specimens cured at RH=60% during the initial drying period (up to 3
days). After the initial drying period, the situation was reversed and specimens cured at
RH=40% showed lower BW. This was likely due to the sharp drying front induced by
SRA (Bentz et al., 2001a, Bentz, 2006), as shown in Fig. 4-9. During early-age, the
initial drying of specimens concentrate SRA in the remaining pore solution at the top
exposed surface of the specimens, which in turns restrict pulling more water from deeper
parts within the specimens (Bentz, 2006). Consequently, higher degree of hydration is
Chapter 4 124
expected at specimen core compared to its surface. The effect of the sharp drying front
seemed to become more important as RH decreased. At later ages, some absorption of
SRA by cement hydration products could reduce the efficiency of the drying front in
eliminating water extraction from the specimen core; leading to a higher rate of
evaporation and a lower BW (Bentz, 2006).
At 10 and 40°C, normal trends were observed. As the RH decreased, the rate of
evaporation increased; hence a lower BW was measured. However, at 40°C a lower
variation in the amount of BW with respect to RH change was observed, as shown in Fig.
4-8(c). This is believed to be due to the lower evaporation rate and the faster development
of the drying front due to the increase in SRA concentration at high temperature, as
mentioned earlier.
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
10°C-60% RH10°C-80% RH10°C-Sealed
(a)
Figure 4-8: Change in BW Vs. ambient humidity for w/c=0.25 + 2% SRA mixture at
a) 10°C, b) 20°C, and c) 40°C.
Chapter 4 125
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
20°C-40% RH20°C-60% RH20°C-80% RH20°C-Sealed
(b)
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
40°C-40% RH40°C-60% RH40°C-80% RH40°C-Sealed
(c)
Figure 4-8 Cont’d: Change in BW Vs. ambient humidity for w/c=0.25 +2% SRA
mixture at a) 10°C, b) 20°C, and c) 40°C.
Chapter 4 126
Figure 4-9: Cross section for SRA specimen exposed to drying condition after 7
days.
Figure 4-10 (a,b,c) shows the effect of RH on the mass loss for w/c=0.25 mixture
incorporating 2% SRA and cured at 10, 20 and 40°C. Generally, a slight reduction in
mass loss compared to that of control mixtures was observed at the different curing
temperatures, which is in agreement with (Weiss and Berke, 2003). However, the mass
loss of SRA mixtures increased with decreasing RH, similar to that of mixtures without
SRA, as previously observed by (Weiss et al., 2008).
At 20°C, the specimens cured at RH=40 and 60% showed similar mass loss
during the initial period of drying. Subsequently, their results started to deviate until the
end of the investigated period, as shown in Fig. 4-10(b). These results agree with earlier
discussion of BW results (Fig. 4- 8(b)). Furthermore, comparing Fig. 4- 10(b) and 4-
10(c), it can be observed that changing the curing temperature did not cause a significant
difference in mass loss. This can also be ascribed to the effect of increasing the curing
temperature on the actual SRA concentration.
Chapter 4 127
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
10°C-60% RH10°C-80% RH
(a)
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
20°C-40% RH20°C-60% RH20°C-80% RH
(b)
Figure 4-10: Change in mass loss Vs. ambient humidity for w/c=0.25 +2% SRA
mixture at a) 10°C, b) 20°C, and c) 40°C.
Chapter 4 128
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
40°C-40% RH40°C-60% RH40°C-80% RH
(c)
Figure 4-10 Cont’d: Change in mass loss Vs. ambient humidity for w/c=0.25 +2% SRA mixture at a) 10°C, b) 20°C, and c) 40°C.
4.5.4.2.2 SAP Mixtures
Figure 4-11(a,b,c) shows the effect of RH levels on the BW for w/c=0.25 mixtures
incorporating 0.6% SAP and cured at 10, 20 and 40°C. Generally, mixtures incorporating
SAP showed a higher degree of hydration, as shown by BW results, compared to those of
mixture without SAP. However, SAP mixtures were highly affected by the RH level.
This was exhibited by higher mass loss when the RH decreased, as shown in Fig. 4-
12(a,b,c). The higher water content due to the additional entrained water held in SAP
particles was likely the main reason for this behaviour. These results are in agreement
with previous findings (Kovler, 1996, Mönnig and Lura, 2007).
Chapter 4 129
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
10°C-60% RH10°C-80% RH10°C-Sealed
(a)
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
20°C-40% RH20°C-60% RH20°C-80% RH20°C-Sealed
(b)
Figure 4-11: Change in BW Vs. ambient humidity for w/c=0.25 +0.6% SAP mixture
at a) 10°C, b) 20°C, and c) 40°C.
Chapter 4 130
0.06
0.07
0.08
0.09
0.10
0.11
0.12
0.13
0.14
0 20 40 60 80 100 120 140 160 180Age (hours)
BW
(g/g
cem
ent)
40°C-40% RH40°C-60% RH40°C-80% RH40°C-Sealed
(c)
Figure 4-11 Cont’d: Change in BW Vs. ambient humidity for w/c=0.25 +0.6% SAP
mixture at a) 10°C, b) 20°C, and c) 40°C.
-6.0-5.5-5.0-4.5-4.0-3.5-3.0-2.5-2.0-1.5-1.0-0.50.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
10°C-60% RH10°C-80% RH
(a)
Figure 4-12: Change in mass loss Vs. ambient humidity for w/c=0.25 +0.6% SAP
mixture at a) 10°C, b) 20°C, and c) 40°C.
Chapter 4 131
-6.0-5.5-5.0-4.5-4.0-3.5-3.0-2.5-2.0-1.5-1.0-0.50.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
20°C-40% RH20°C-60% RH20°C-80% RH
(b)
-6.0-5.5-5.0-4.5-4.0-3.5-3.0-2.5-2.0-1.5-1.0-0.50.0
0 20 40 60 80 100 120 140 160 180Age (hours)
Mas
s lo
ss(%
)
40°C-40% RH40°C-60% RH40°C-80% RH
(c)
Figure 4-12 Cont’d: Change in mass loss Vs. ambient humidity for w/c=0.25 +0.6%
SAP mixture at a) 10°C, b) 20°C, and c) 40°C.
Chapter 4 132
Due to higher ion concentration inside SAP, water flows under the effect of
osmosis from the cement matrix to SAP particles (Jensen and Hansen, 2001). As the
hydration progresses, more ions dissolve in the pore solution (Lea, 1988). Once the ions
concentration in the pore solution becomes higher than that inside SAP particles, water
starts to flow in the opposite direction (i.e. from SAP particles to the surrounding cement
matrix (Jensen and Hansen, 2001)). Combined with water loss due to evaporation, this
resulted in higher ions concentration in the pore solution, which motivates further water
flow from the SAP particles. Furthermore, since the rate of hydration and evaporation are
functions of the temperature and RH, it is expected that increasing the curing temperature
and/or decreasing the RH will result in more extracted water from the SAP particles, thus
enhancing evaporation and higher mass loss.
4.5.5. Correlation between Autogenous Strain and Achieved Degree of Hydration
under Sealed Condition
4.5.5.1 SRA Mixtures
Autogenous strains are shown in Fig. 4-13 as a function of the RBW for mixtures
incorporating 2% SRA and cured at 10, 20 and 40°C. A good correlation can be observed
between the measured autogenous strains and the RWB for mixtures cured at 10 and
20°C. At 40°C, a different trend was observed. At the same RBW, a higher autogenous
strain was achieved at 40°C compared to that at 10 and 20°C. This could be ascribed to
the absence of early expansion at 40°C due to the adverse effect of the high curing
temperature at early-age on ettringite formation and the acceleration of hydration, leading
to higher self-desiccation. Mono-sulphate super-saturation levels and stability (Thomas et
Chapter 4 133
al., 2003) will be affected, leading to lower rate of ettringite formation. Moreover, the
higher self-desiccation shrinkage opposes expansion induced by any ettringite formation
(Bentz et al., 2001b, Kaufmann et al., 2004).
(R2 = 0.98)
-800
-700
-600
-500
-400
-300
-200
-100
00 0.02 0.04 0.06 0.08 0.1
RBW (g/g cement)
Aut
ogen
ous
Stra
in (µ
ε)
w/c=0.22 - 10°Cw/c=0.22 - 20°Cw/c=0.22 - 40°Cw/c=0.25 - 10°Cw/c=0.25 - 20°Cw/c=0.25 - 40°C
40°C
10&20°
Figure 4- 13: Correlation between autogenous strains and RBW for 2% SRA
mixtures cured at 10, 20 and 40°C.
4.5.5.2 SAP Mixtures
For mixtures incorporating 0.6% SAP, experimental results revealed a linear correlation
between the measured autogenous strains and RBW at 20 and 40°C, as shown in Fig. 4-
14. At 10°C, no correlation was observed, which can be ascribed to the effect of the high
early-age expansion due to a disjoining pressure provided by water released from the
SAP. Such an expansion, depending on its magnitude and rate of development, will act to
offset deformations due to shrinkage. Further research is needed to explore this complex
combination of expansion and autogenous strain. This includes, for instance, the
Chapter 4 134
influence of the rate of dissolution of salts in the pore solution on capillary stresses, the
disjoining pressure (Lura, 2003), and the coefficient of super-saturation at low
temperature for mixtures incorporating SAP.
R2 = 0.97
R2 = 0.55
-800
-700
-600
-500
-400
-300
-200
-100
00 0.02 0.04 0.06 0.08 0.1
RBW (g/g cement)
Aut
ogen
ous
Stra
in (µ
ε)
w/c=0.22 - 10°Cw/c=0.22 - 20°Cw/c=0.22 - 40°Cw/c=0.25 - 10°Cw/c=0.25 - 20°Cw/c=0.25 - 40°C
10°C
20&40°C
Figure 4- 14: Correlation between autogenous strains and RBW for 0.6% SAP
mixtures cured at 10, 20 and 40°C.
4.5.6. Predicting Drying and Autogenous Strains under Different Curing
Conditions
To investigate the effect of drying conditions on the role of SRA and SAP, the drying
strains were determined based on the superposition principle and the evaluated degree of
hydration (described in Chapter 3).
Chapter 4 135
4.5.6.1 SRA Mixtures
Similar to the control mixtures, the contribution of autogenous strains to the total strain
under drying conditions for mixtures incorporating 2% SRA, based on the superposition
principle, were overestimated when compared with that obtained using the degree of
hydration, as shown in Fig. 4-15. However, SRA mixtures showed lower variation
between the evaluated autogenous strains under different RH and sealed specimens
compared to that of the control mixtures (lower overestimation values (Table 4-1)). For
instance, at 40°C the difference between autogenous strains at RH=40% and sealed
specimens was about 50 and 360 µε for SRA and control mixtures, respectively. This can
be ascribed to the role of SRA in maintaining higher internal relative humidity under
different drying conditions, which enhanced the hydration process (Bentz et al., 2001a,
Weiss et al., 2008). SRA effectively reduced the drying strains for specimens cured under
low RH conditions, while reduced autogenous strain slightly. For instance, w/c=0.25
+2% SRA specimens achieved up to 55% reduction in the drying strain at 40°C and RH=
40% compared to only 6% reduction in the autogenous strain at 40°C. Conversely, a
higher efficiency of SRA in mitigating autogenous strains was observed at high RH
compared to that of the control mixtures, as shown in Fig. 4-16. In conclusion, SRA was
more effective in reducing drying strains at low RH, where drying dominated the total
deformation and autogenous deformation had a minor effect. It can also be argued that
SRA can be more effective in reducing autogenous strains for sealed or large cross-
section concrete members, where autogenous deformation at the core would be the major
cause of deformation.
Chapter 4 136
-1000
-800
-600
-400
-200
0
Aut
ogen
ous
stra
in (µ
ε)
10°C20°C40°C
RH=40% RH=80%RH=60% SealedAmbient humidity
Figure 4-15: Autogenous strains evaluated based on achieved degree of hydration
for w/c=0.25 + 2% SRA UHPC mixtures at age 7 days.
-1000
-800
-600
-400
-200
0
Aut
ogen
ous
Stra
in (µ
ε)
40°C-Control40°C-SRA20°C-Control20°C-SRA10°C-Control10°C-SRA
RH=40% RH=80%RH=60%Ambient Humidity
Figure 4-16: Comparison between autogenous strains evaluated based on achieved
degree of hydration for w/c=0.25 mixtures at age of 7-days with and without SRA.
Chapter 4 137
Table 4- 1: Overestimation ratios for autogenous strains under different curing
conditions for different w/c=0.25 UHPC mixtures
Overestimation ratio
Mixture Temp.°C/RH% 40 60 80
10 ‐‐‐‐* 0.78 0.46
20 1.26 0.98 0.45 Control
40 1.74 1.17 0.63
10 ‐‐‐‐* 0.47 0.15
20 0.57 0.46 0.26 2% SRA
40 0.32 0.16 0.13
20 0.98 0.47 0.20 0.6% SAP
40 1.13 0.67 0.48
* Condition is not feasible.
4.5.6.2 SAP Mixtures
Figure 4-17 shows autogenous strains evaluated based on the degree of hydration for
mixtures incorporating 0.6% SAP under different exposure conditions (except at 10°C,
no correlation between the strains and degree of hydration was obtained as discussed
earlier). It can be observed that autogenous strains for sealed specimens were
significantly higher than that under other exposure conditions as shown by the
overestimation ratio in Table 4-1. In addition, SAP efficiency in mitigating autogenous
strains decreased with lowering the RH. Lowering the RH facilitates water escaping
from specimens, which disturbs water storage in the SAP particles as mentioned before.
This can affect the role of SAP in mitigating autogenous strains induced by self-
desiccation. For instance, the reduction in autogenous strains at 40°C for SAP mixtures
Chapter 4 138
compared to that of control mixtures was about 7 and 21% for RH=40 and 80%,
respectively as shown in Fig. 4-18. Furthermore, SAP mixtures exhibited higher drying
strains compared to that of the control mixtures which is in agreement with mass loss
results (see Fig. 4-12(a,b,c)).
-1000
-800
-600
-400
-200
0
Aut
ogen
ous
stra
in (µ
ε)
20°C
40°C
RH=40% RH=80%RH=60% SealedAmbient humidity
Figure 4-17: Autogenous strains evaluated based on achieved degree of hydration
for w/c=0.25 + 0.6% SAP UHPC mixtures at age 7 days.
Chapter 4 139
-1000
-800
-600
-400
-200
0
Aut
ogen
ous
stra
in (µ
ε)
40°C-Control40°C-SAP20°C-Control20°C-SAP
RH=40% RH=80%RH=60%Ambient humidity
Figure 4-18: Comparison between autogenous strains evaluated based on achieved
degree of hydration for w/c=0.25 mixtures at age of 7-days with and without SAP.
4.6. CONCLUSIONS
Using the amount of chemically bound water to assess autogenous and drying strains in
ultra high-performance concrete specimens was explored in this study including the
effects of a shrinkage reducing admixture and superabsorbent polymer. The following
conclusions can be drawn from the experimental results:
1) The reduction in UHPC compressive strength induced by the addition of SRA and
SAP should be considered in evaluating the overall field performance.
2) Effect of different shrinkage mitigation techniques on development of autogenous
and drying strains are dependent; therefore their behaviour under sealed
conditions will differ from that under drying conditions.
Chapter 4 140
3) Applying the superposition principle without considering the effect of drying
conditions on the shrinkage mitigation mechanisms will result in overestimating
autogenous strains, and consequently an unrealistic evaluation for the efficiency
of the applied shrinkage mitigation technique.
4) For thin UHPC sections, drying dominates the total deformation and reduces the
development of autogenous strains, leading to lower autogenous contribution to
the total deformation.
5) Adding SRA effectively reduced drying strains, which are the dominant strains in
UHPC specimens under low RH conditions. At higher RH conditions, SRA
reduced autogenous strains, which in turn are the dominant strains at high RH.
6) In sealed UHPC specimens, early-age expansion of SAP mixtures had a
significant effect in reducing the net strains.
7) In UHPC specimens under drying conditions, adding SAP resulted in higher
drying strains, which disturbed the curing process and diminished the effect of
SAP as a shrinkage mitigating method.
8) Adequate external curing is essential to mitigate early-age deformation in UHPC
even when internal curing mechanisms are provided, since it guarantees a suitable
environment for shrinkage mitigation methods to work properly.
Chapter 4 141
4.7. REFERENCES
Bao-guo, M., Xiao-dong, W., Ming-yuan, W., Jia-jia, Y. and Gao, X-j., (2007), “Drying Shrinkage of Cement-Based Materials Under Conditions of Constant Temperature and Varying Humidity,” Journal of China Univiversity for Mining and Technology, Vol. 17, No. 3, pp. 428-431.
Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafai, N., (2006),
“Autogenous deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp. 123-136.
Bentz, D.P., (2006), “Influence of Shrinkage-Reducing Admixtures on Early-Age
Properties of Cement Pastes,” Journal of Advanced Concrete Technology, Vol. 4, No. 3, 423-429.
Bentz, D.P., Geiker, M.R. and Hansen, K.K., (2001a), “Shrinkage-reducing admixtures
and early-age desiccation in cement pastes and mortars,” Cement and Concrete Research, Vol. 31, No. 7, pp. 1075-1085.
Bentz, D.P., Jensen, O.M., Hansen, K.K., Olesen, J.F., Stang, H. and Haecker, C-J.,
(2001b), “Influence of cement particle-size distribution on early age autogenous strains and stresses in cement-based materials,” Journal of American Ceramic Society, Vol. 84, No. 1, pp. 129-135.
Bentz, D.P. and Peltz, M.A., (2008), “Reducing Thermal and autogenous shrinkage
contributions to early-age cracking,” ACI Materials Journal, Vol. 105, No. 4, pp. 414-420.
Cusson, D. and Hoogeveen, T.J., (2007), “An experimental approach for the analysis of
early-age behaviour of high-performance concrete structures under restrained shrinkage,” Cement and Concrete Research, Vol. 37, No. 2, pp. 200-209.
D’Ambrosia, M., (2002), “Early-age Tensile Creep and Shrinkage of Concrete with
Shrinkage Reducing Admixtures,” Master Thesis, University of Illinois at Urbana-Champaign, United States.
Jensen, O.M. and Hansen, P.F., (2001), “Water-entrained cement-based materials: I.
Principles and theoretical background,” Cement and Concrete Research, Vol. 31, No. 4, pp. 647-654.
Jensen, O.M. and Hansen, P.F., (2002), “Water-entrained cement-based materials: II. Experimental observations,” Cement and Concrete Research, Vol. 32, No. 6, pp. 973-978.
Kamen, A., Denarié, E., Sadouki, H. and Brühwiler, E., (2008), “Thermo-mechanical
response of UHPFRC at early age - Experimental study and numerical simulation,” Cement and Concrete Research, Vol. 38, No. 6, pp. 822-831.
Chapter 4 142
Kaufmann, J., Winnefeld, F. and Hesselbarth, D., (2004), “Effect of the addition of
ultrafine cement and short fiber reinforcement on shrinkage, rheological and mechanical properties of Portland cement pastes,” Cement and Concrete Composites, Vol. 26, No. 5, pp. 541-549.
Kovler, K., (1996), “Why Sealed Concrete Swells,” ACI Materials Journal, Vol. 93, No.
4, pp. 334-340. Lam, H., (2005), “Effects of internal curing methods on restrained shrinkage and
permeability,” Master Thesis., University of Toronto Lea, F.M., (1988). Lea's Chemistry of Cement and Concrete, 4th Edition., J. Wiley, New
York. Loser, R. and Leemann, A., (2009), “Shrinkage and restrained shrinkage cracking of self-
compacting concrete compared to conventionally vibrated concrete,” Materials and Structures, Vol. 42, No. 1, pp. 71-82.
Lura, P., (2003), “Autogenous deformation and internal curing of concrete,” PhD. Thesis,
Technical University Delft, Netherlands. Lura, P. and Jensen, O.M., (2007), “Measuring techniques for autogenous strain of
cement paste,” Materials and Structures, Vol. 40, No. 4, pp. 431-440. Mikhail, R.S., Abo-El-Enein, S.A. and Gabr, N.A., (1977), “Hardened slag-cement pastes
of various porosities I. Compressive strength, degree of hydration and total porosity,” Journal of Applied Chemistry and Biotechnology, Vol. 24, No. 12, pp. 735-743.
Mönnig, S. and Lura, P., (2007), “Superabsorbent Polymers - An Additive to Increase the Freeze-Thaw Resistance of High Strength Concrete,” In: Advances in Construction Materials, Part V, Springer Berlin Heidelberg, pp. 351-358.
Mounanga, P., Khelidj, A., Loukili, A. and Baroghel-Bouny, V., (2004), “Predicting
Ca(OH)2 content and chemical shrinkage of hydrating cement pastes using analytical approach,” Cement and Concrete Research, Vol. 34, No. 2, pp.255-265.
Odler, I. and Rößler, M., (1985), “Investigations on the relationship between porosity,
structure and strength of hydrated Portland cement pastes: II. Effect of pore structure and of degree of hydration,” Cement and Concrete Research, Vol. 15, No. 3, pp. 401-410.
Rajabipour, F., Sant, G. and Weiss, J., (2008), “Interactions between shrinkage reducing
admixtures (SRA) and cement paste's pore solution,” Cement and Concrete Research, Vol. 38, No. 5, pp. 606-615.
Chapter 4 143
Thomas, J.J., Rothstein. D., Jennings, H.M. and Christensen, B.J., (2003), “Effect of hydration temperature on the solubility behavior of Ca-, S-, Al-, and Si-bearing solid phases in Portland cement pastes,” Cement and Concrete Research, Vol. 33, No. 12, pp. 2037-2047.
Wang, F., Zhou, Y., Peng, B., Liu, Z. and Hu, S., (2009), “Autogenous shrinkage of
concrete with super-absorbent polymer,” ACI Materials Journal, Vol. 106, No. 2, pp. 123-127.
Weiss, J. and Berke, N., (2003), “Admixtures for reduction of shrinkage and cracking,”
In: Early Age Cracking in Cementitious Systems: Report of RILEM Technical Committee 181-EAS, Bentur A; (ed.), pp. 323-335.
Weiss, J., Lura, P., Rajabipour F. and Sant. G., (2008), “Performance of shrinkage-
reducing admixtures at different humidities and at early ages,” ACI Materials Journal, Vol. 105, No. 5, pp. 478-486.
Yuasa, N., Kasai, Y., Matsui, I., (1999), “Inhomogeneous distribution of the moisture
content and porosity in concrete,” In: Proceedings of the International Conference, R.K. Dhir and M.J. McCarthy, Editors, the University of Dundee, Scotland, United Kingdom, Thomas Telford Publ., pp. 93-101.
Chapter 5 144
CHAPTER FIVE
EARLY-AGE SHRINKAGE OF UHPC UNDER DRYING/WETTING CYCLES AND SUBMERGED
CONDITIONS*
In the previous chapters (3 and 4), the influence of coupling a wide range of temperatures
and relative humidities on drying and autogenous shrinkage behaviour, and their
interaction mechanisms was investigated. Moreover, the role of shrinkage reducing
admixture (SRA) and superabsorbent polymer (SAP), as shrinkage mitigation techniques,
and their efficiency were also explored under similar conditions. However, such tests still
not capturing the effect of other factors existing in field conditions, which may modify
the early-age shrinkage behaviour of UHPC and its interaction with surrounding
environment. Therefore, the present chapter examines the early-age shrinkage behaviour
of UHPC mixtures, with and without shrinkage mitigation techniques, along with
considering environmental loads (including the application of drying/wetting cycles
and/or full immersion). This shall initiate the idea for developing a comprehensive
integrated early-age shrinkage testing approach.
5.1. INTRODUCTION
Concrete is usually subjected to early-age variation in temperature, relative humidity,
wind and other environmental loading that can alter its behaviour. Considering in-situ
conditions and environmental loading in testing protocols should allow gaining a better
*A version of this chapter has been Accepted in the ACI Materials Journal. Some parts of this chapter were also published in the Second International Structures Specialty Conference, Winnipeg, Manitoba, Canada, ST-30.
Chapter 5 145
understanding of the early-age behaviour and developing suitable performance
specifications. This seems to be more important for new generations of concrete (i.e.
UHPC) before being implemented in wide full-scale field constructions.
During early-age curing, before depercolation of capillary pores, concrete can
imbibe water from its environment. Hence, self-desiccation shrinkage can be delayed and
cracking may be avoided (Bentz and Jensen, 2004). In addition, early curing minimizes
early-age moisture loss and maximizes the degree of hydration achieved by the cement,
potentially leading to stronger and more durable concrete. Based on the quality of the
curing regime, UHPC can be subjected to cyclic drying (intervals between exposures to
moisture) and wetting (during moisture exposure). Besides the curing regime, concrete is
usually exposed in the field to various environmental conditions. These may include: the
drying action of the sun and wind; rain-water containing dissolved chemicals; bridge-
deck run-off or road spray contaminated with chlorides from de-icing salts; tidal and
wave action in offshore or water-front marine structures, etc. Hence, it can be expected
that the quality of the curing regime and the effects of the surrounding environment may
be responsible for considerable dimensional instability due to moisture changes, and can
thus have major consequences on the long-term performance of concrete.
Therefore, the present chapter explores the shrinkage behaviour of UHPC under
different environmental loads. These loads include drying/wetting (DW) cycles, which
used as an accelerated test method to simulate outdoor environmental conditions in which
structures are subjected to moisture cycles, and submerged (SM) condition to simulate the
submerged part of marine and offshore structures.
Chapter 5 146
5.2. RESEARCH SIGNIFICANCE
Several studies have been conducted to quantify the shrinkage behaviour of UHPC using
a conventional testing approach under controlled environment. However, achieving a
proper understanding for the effect of field exposure conditions on such shrinkage a more
integrated testing approach is still needed. Hence, this study introduces a simple
integrated testing method through considering the effect of environmental load on
shrinkage behaviour and the efficiency of shrinkage mitigation methods. Results should
have important implications in understanding the evolution of early-age deformations in
UHPC and the role of mitigation methods under field-like conditions. In addition, it
paves the way for developing a multi-factor integrated testing approach for early-age
shrinkage behaviour that can capture different field scenarios.
5.3. EXPERIMENTAL PROCEDURE
5.3.1. Materials and Mixture Proportions
The materials, UHPC mixtures and ID coding used in this chapter were similar to that
used in Chapter 4 (refer to Section 4.4.1). A new UHPC mixture incorporating 2% SRA
and 0.6% SAP was tested in order to investigate the synergetic effect of both mitigation
techniques. The chemical and physical properties of the used binders have been given in
Chapter 3 (Tables 3-1).
5.3.2. Environmental Conditions
Four exposure conditions were simulated, namely; drying, sealed, SM and DW cycles.
The drying condition was simulated inside a walk-in environmental room, where
Chapter 5 147
temperature and relative humidity (RH) were kept constant (20 ±1°C and RH= 40 ± 5%).
Submerged specimens were stored in a temperature controlled water bath at 20 ±2°C.
During the DW, specimens were allowed to dry inside the walk-in environmental room at
20 ±1°C and RH= 40% for 12 hours, then submerged in a temperature controlled water
bath at 20 ± 2°C for 6 hours. Controlling the temperature of specimens insures a similar
temperature history; hence, the temperature effects on the developed autogenous
shrinkage strains are minimized during the various exposure regimes (Turcry et al.,
2002).
5.3.3. Preparation of Test Specimens and Testing Procedures
Prismatic specimens (75×75×280 mm) were used for all conducted tests in order to dispel
size effects on the results. All UHPC specimens were taken from a single batch of the
tested mixture. Before casting, the moulds were lined with a layer of thin Teflon sheets to
minimize friction between the concrete and moulds. Specimens were cast in three layers
and compacted on a vibrating table. UHPC was cast in metallic molds with 5-mm thick
copper walls. Silicone sealant was used to seal the molds’ sides and prevent any water
diffusion from the bath. The surface of each specimen was covered with a wide
polyethylene sheet which was bent down on both sides of the mold and fixed with a
silicone sealant. The temperature of each specimen was controlled using a temperature
controlled water bath surrounding the specimens’ sides. The top surface temperature was
controlled using a wet sponge. The sponge was submerged in the water bath regularly
every 15 minutes until the demolding time. At around 5-7 hours from first contact
between cement and water, molds were removed from the water bath. The silicon sealant
and wall-base attaching screws were removed, then specimens were taken off and moved
Chapter 5 148
to the specific curing conditions. The specimens were installed over roller supports that
can rise up the specimens allowing: i) drying from all specimens’ faces, and ii) specimens
to shrink freely without friction effects. In the remainder of this text, strains will be used
to account for both shrinkage and expansion deformations.
5.3.3.1 Chemically Bound Water and Degree of Hydration
Thermo-gravimetric (TGA) combined with derivative thermo-gravimetric (DTG)
analysis was previously explained in Chapter 3 (refer to section 3.4.3.1)
5.3.3.2 Strain Measurements
The strain measurements on UHPC specimens were carried out using electrical-resistance
strain gauges according to a common procedure (RILEM, 2003, Tazawa, 1999) in which
an embedded steel bar, with strain gauges installed on it, is used to measure strain inside
the tested specimen. Fresh UHPC mixtures were cast in metallic formwork with steel
reinforcing bars (6 mm in diameter, and 280 mm in length). On each steel bar, four strain
gauges were installed to form a full bridge circuit, as shown in Fig. 5-1. Two strain
gauges were installed opposite to each other along a fabricated 20 mm smooth length at
the middle of the bar to eliminate the bending effect on the measured strains (Hannah and
Reed, 1992, Sule and Van Breugel, 2001). The other two strain gauges were installed in
the transverse direction to minimize the temperature effect on the measured strains
(RILEM, 2003, Hannah and Reed, 1992). In addition, two metallic plates were added at
the ends of the bar to capture any expansion deformation. The measured strains were
corrected to account for the local restraint induced by the embedded bar due to the
difference between the stiffness of concrete and that of the embedded steel bar as follows
(RILEM, 2003):
Chapter 5 149
TAEAE
scc
ssmt ∆+⎥
⎦
⎤⎢⎣
⎡+= αεε 1 Eq. 5-1
Where εt is the total concrete strain after correction, εm is the measured strain after
deducting the temperature effect on the measurements, Es, As and Ec, Ac are the modulus
of elasticity and cross-section area of the embedded steel bar and tested UHPC
specimens, respectively, αs is the coefficient of thermal expansion of the embedded steel
bar, and T∆ is the change in temperature from the onset of the test. The calculated value
(Es As /Ec Ac) was much lower than 1, which indicates a very low restraint induced by the
steel bar against shrinkage.
Figure 5-1 : (a) Strain measurement specimen, and (b) strain gauges arrangement to
form full bridge circuit.
For each UHPC mixture, three replicate specimens per curing condition were
made. Immediately after demolding, specimens for autogenous strain measurements were
wrapped with four layers of polyethylene sheets and a layer of paraffin wax membrane to
prevent moisture loss. Specimens for drying and drying/wetting strain measurements
were exposed to the drying condition inside the walk-in environmental room. Submerged
specimens were moved to the controlled water bath. Moreover, two type-T
Chapter 5 150
thermocouples were inserted in the specimens to monitor concrete temperature changes
from the onset of testing.
5.3.3.3 Moisture Loss
Three identical prisms (75×75×280 mm) were made for each mixture and demolded at
the time of starting strain measurements. The prisms were transferred to the different
exposure conditions after measuring the initial mass of each prism using a balance with
an accuracy of 0.01 g. The mass measurements were started along with measurements of
total strains.
5.3.3.4 Mercury Intrusion Porosimetry (MIP):
UHPC fragments were taken from tested specimens and immediately plunged in an
isopropanol solvent to stop hydration and subsequently dried inside a desiccator until a
constant mass was achieved. The pore size distribution for each specimen was
determined automatically using a Micromeritics AutoPore IV 9500 Series porosimeter
allowing a range of pressures from 0 to 414 MPa. The assumed surface tension of
mercury was about 0.484 N/m at 25°C according to ASTM D 4404-84 (Standard Test
Method for Determination of Pore Volume and Pore Volume Distribution of Soil and
Rock by Mercury Intrusion Porosimetry). The density of the mercury was 13.546 g/ml
and the assumed contact angle was 140°.
5.3.3.5 Chemical Oxygen Demand (COD)
Chemical Oxygen Demand (COD) test determines the oxygen requirement equivalent of
organic matter that is susceptible to oxidation with the help of a strong chemical oxidant.
In the COD method, the water sample is oxidized by digesting in a sealed reaction tube
Chapter 5 151
with sulphuric acid and potassium dichromate in the presence of a silver sulphate
catalyst. The amount of dichromate reduced is proportional to the COD. A reagent blank
is prepared for each batch of tubes in order to compensate for the oxygen demand of the
reagent itself. Over the range of the test a series of colours from yellow through green to
blue are produced. The colour is indicative of the chemical oxygen demand and is
measured using a Photometer. The results are expressed as milligrams of oxygen
consumed per litre of sample.
5.4. RESULTS AND DISCUSSION
5.4.1. Mass Change
The mass change of the control UHPC specimens due to the ingress/absorption of
moisture is illustrated in Fig. 5-2. Under the drying condition, specimens showed a
continuous mass loss over the investigated period. Conversely, specimens subjected to
DW cycles showed fluctuation of mass change while switching between the drying and
wetting periods. On the other hand, submerged specimens gained mass rapidly during the
first 24 hours, and then their mass became nearly constant. This can be attributed to the
higher ability for imbibing water during the early period before depercolation of capillary
porosity (Bentz and Garboczi, 1991). However, control specimens with w/c=0.22
exhibited a lower mass change of about 25% and 58% under the drying and submerged
conditions, respectively, compared to that of control specimens with w/c=0.25. This can
be ascribed to the reduction in porosity as a result of reducing the w/c. Reducing the w/c
generally leads to lower capillary suction of water and reduced diffusion coefficient due
to the development of a denser pore structure (Lea and Hewlett, 1998, Martys and
Chapter 5 152
Ferraris, 1997, Teichmann and Schmidt, 2004). Moreover, the w/c=0.22 control mixture
had a lower content of evaporable water compared to that of the w/c=0.25 control
mixture due to its lower water content and higher consumption of evaporable water in the
hydration process (i.e. self-desiccation) (Dhir and McCarthy, 1999).
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0 24 48 72 96 120 144 168
Age (hours)
Mas
s ch
ange
(%)
DryDrying/WettingSubmerged
Submerged
Drying/wetting
Dry
(a)
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0 24 48 72 96 120 144 168
Age (hours)
Mas
s ch
ange
(%)
DryDrying/WettingSubmerged
Submerged
Dry
Drying/wetting
(b)
Figure 5-2: Mass change of control mixtures a) w/c=0.22 and b) w/c=0.25 under
different exposure conditions.
Chapter 5 153
Similar trends were observed for UHPC mixtures incorporating SRA and SAP
(Fig. 5-3). Compared with control mixtures, the SRA mixtures did not exhibit a
significant mass change under drying and DW conditions. Conversely, the SAP mixtures
showed a higher mass loss. Under the SM condition, SRA and SAP mixtures behaved
differently when the w/c changed. SRA and SAP mixtures with w/c=0.25 showed a lower
mass gain compared to that of the w/c=0.25 control mixture. This can be ascribed to the
effect of SRA and SAP on the capillary suction, which is considered the driving force for
water diffusion into the porous material (Lstiburek and Carmody, 1994). Using SRA is
believed to reduce the capillary suction since it reduces the pore fluid surface tension
(Ribeiro et al., 2006). The SAP mixture had a higher water content leading to lower
capillary suction (Lstiburek and Carmody, 1994). On the other hand, SRA and SAP
mixtures with a w/c=0.22 showed similar mass gain compared to that of the w/c=0.22
control mixture (see Fig. 5-3), which indicates that the water diffusion and consequently
mass change at such a low w/c is dominated by the development and depercolation rate
of the pore structure leading to lower diffusion characteristics (Lea and Hewlett, 1998).
Chapter 5 154
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Drying Drying/Wetting Submerged
Mas
s ch
ange
(%)
ControlSRASAP
(a)
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Drying Drying/Wetting Submerged
Mas
s ch
ange
(%)
ControlSRASAP
(b)
Figure 5-3: Mass change for a) w/c=0.22 and b) w/c= 0.25 mixtures incorporating
2% SRA or 0.6% SAP under different exposure conditions.
Chapter 5 155
5.4.2. Shrinkage Strain under Different Environmental Exposure Conditions
5.4.2.1 Control Mixture
UHPC specimens exposed to the drying and sealed conditions showed significant
shrinkage strain. DW cycles seemed to be somewhat effective in counteracting shrinkage
strain. The final shrinkage strains after DW cycles were reduced by about 21% and 37%
compared to that under the drying condition for mixtures with w/c=0.22 and 0.25,
respectively. Figure 5-4 shows the strain development for the w/c=0.25 control mixture
under different curing conditions.
-1000
-800
-600
-400
-200
0
200
0 24 48 72 96 120 144 168
Age (hours)
Stra
in (µ
ε)
DryAutogenousDrying/wettingSubmerged
Submerged
Dry
Autogenous
Drying/wetting
Figure 5-4: Measured strain for w/c=0.25 control specimens under different
exposure conditions.
It can be seen that during the wetting period in DW cycles (Fig. 5-5), specimens
expanded as a result of water absorption. Conversely, during the drying period, it started
to shrink due to the loss of moisture, along with thermal deformation caused by
Chapter 5 156
evaporative cooling due to the removal of free water (Kovler, 1996). These thermal
deformations (which ranged from 20 to 35 µε) were eliminated from strain measurements
based on the measured specimens’ temperature and coefficient of thermal expansion,
which was evaluated using a similar procedure to that proposed by (Cusson and
Hoogeveen, 2006, Cusson and Hoogeveen, 2007).
-600
-580
-560
-540
-520
-500
-480
-460
0 24 48 72 96 120 144 168Age (hours)
Stra
in (µ
ε)
Drying cycle
Wetting cycle
Figure 5-5: Measured strain for w/c=0.25 control specimens under drying/ wetting
cycles.
Moreover, the amount and rate of strain development during either the drying or
wetting period was found to decrease with time. This can be explained as follows: the
higher the pore volume of the specimen, the deeper was the level of internal free water
that can evaporate and be exchanged with the surrounding environment (Aïtcin, 1998),
the so-called moisture influential depth (Chunqiu et al., 2008). As the hydration process
progresses, especially for UHPC, hydration products start filling voids and pores, thus
Chapter 5 157
leading to denser internal microstructure (Loukili et al., 1999, Lstiburek and Carmody,
1994). This will limit the exchange of water with the external surface layer of the
specimen. Figure 5-6(a,b) illustrate the progress of the degree of hydration and pore
volume for the surface layers and cores of UHPC specimens with w/c=0.25 subjected to
DW cycles. It can be observed that the cores of specimens have higher degree of
hydration and denser pore structure compared to that of the external layer.
6
8
10
12
14
16
18
0 24 48 72 96 120 144 168Age (hours)
Poro
sity
(%)
Submerged (S)Submerged (C)Dry (S)Dry (C)AutogenousDrying/wetting (S)Drying/Wetting (C)
(a)
Figure 5-6: Measured (a) porosity and (b) degree of hydration for w/c=0.25 control
specimens surface (S) and core (C) under different exposure conditions.
Chapter 5 158
30
34
38
42
46
50
0 24 48 72 96 120 144 168Age (hours)
Deg
ree
of h
ydra
tion
(%)
Submerged (S) Submerged (C)Dry (S) Dry (C)Autogenous Drying/Wetting (S)Drying/Wetting (C)
(b)
Figure 5-6 Contd’: Measured (a) porosity and (b) degree of hydration for w/c=0.25
control specimens surface (S) and core (C) under different exposure conditions.
Under SM conditions, specimens continued to swell during the first 24 hours.
Then, it started to shrink until the end of the investigated period. This high early swelling
resulted in a lower net shrinkage compared with that of sealed specimens (about 79% and
98% reduction in the shrinkage strain for mixtures with w/c=0.22 and 0.25, respectively).
The swelling of submerged specimens can be ascribed to the continuous supply of water,
allowing the cement gel to absorb water and expand (Neville, 1996). The reduction in the
swelling strain can be explained by the fact that the progress of the hydration process
reduced and depercolated the concrete capillary porosity, which interfered with water
imbibing from the surrounding environment to the specimen’s core. As a result, higher
self-desiccation occurred as water was drained from increasingly finer capillaries, thus
leading to higher capillary tension and higher deformations. Unlike the case of DW
Chapter 5 159
cycles, the external layers were found to be denser and had higher degree of hydration
than that of specimen cores as shown in Fig. 5-6 (a,b).
However a similar behaviour was observed for the w/c=0.22 control mixture,
which showed a lower rate and amount of strain during DW cycles and under the SM
condition compared to that of the w/c=0.25 control mixture. For instance, the w/c=0.22
control mixture exhibited an early expansion strain during the first 24 hours of about 21%
less than that of the w/c=0.25 control mixture. This behaviour of the w/c=0.22 mixture
can be attributed to its denser pore structure, which eliminated the water exchange
between the specimens and their surrounding (Tazawa, 1999), whether during drying or
wetting periods. Figure 5-7 illustrates the effects of the w/c and exposure conditions on
the surface pore volume distribution after one DW cycle. It can be observed that reducing
the w/c and applying moist curing resulted in lower total porosity values and movement
of the pore width peak to smaller widths in agreement with previous research (Cook and
Hover, 1999, Alford and Rahman, 1981).
Chapter 5 160
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
1 10 100 1000 10000 100000Pore diameter (nm)
dV/d
Log(
D) (
ml/g
.nm
)
0
1
2
3
4
5
6
7
dV/d
Log(
D) (
cu.in
/oz.
µin)
w/c=0.22-Dryw/c=0.22-Drying/wettingw/c=0.22-Submergedw/c=0.25-Dryw/c=0.25-Drying/wettingw/c=0.25-Submerged
Figure 5-7: Pore size distribution for control specimen’s surface under different
exposure conditions after the first drying/wetting cycle.
5.4.2.2 Effect of SRA
Figure 5-8 illustrates the shrinkage strain for w/c=0.25 specimens incorporating SRA
under different curing conditions. Incorporating SRA led to a significant reduction in
both the drying and autogenous strains by about 24 and 38% for w/c=0.22 mixtures and
about 18% and 25% for w/c= 0.25 mixtures, respectively, compared to that of
corresponding control specimens with no SRA. This is attributed to the effect of SRA in
reducing the surface tension of the pore fluid, leading to a reduction in the developed
capillary stress and drying shrinkage strains (Bentz, 2006, Weiss et al., 2008). Moreover,
SRA is believed to mitigate the drop in internal relative humidity of specimens, leading
to lower self-desiccation and autogenous shrinkage (Bentz et al., 2001).
Chapter 5 161
-1000
-800
-600
-400
-200
0
200
400
0 24 48 72 96 120 144 168
Age (hours)
Stra
in (µ
ε)
DryAutogenousDrying/wettingSubmerged
Submerged
Drying/wetting
AutogenousDry
Figure 5-8: Measured strain for w/c=0.25 UHPC specimens incorporating 2% SRA
under different exposure conditions.
Figure 5-9 illustrates changes in the drying strain rate during DW cycles for the
control and SRA mixtures. Two interesting features can be observed. First, a steep
reduction in the drying strain rate after about 1 and 3 DW cycles is observed for w/c=0.22
and 0.25 mixtures, respectively. As hydration proceeds, hydration products fill the pore
space, thus diminishing the size and connectivity of pores (Cook and Hover, 1999). The
resulting lower water exchange with the surrounding environment causes a steep
reduction in the drying strain rate. The lower the w/c, the lower was the pore space to be
filled, leading to earlier reduction in the rate of water exchange (Loukili et al., 1999,
Lstiburek and Carmody, 1994).
Chapter 5 162
0.5
1.0
1.5
2.0
2.5
3.0
1 2 3 4 5 6 7 8 9Number of drying/wetting cycles
Dry
ing
stra
in ra
te (
µε/h
r)
Control w/c=0.22SRA w/c=0.22Control w/c=0.25SRA w/c=0.25
Figure 5-9: Drying strain rate during drying/wetting cycles for control and SRA
UHPC specimens.
Secondly, mixtures incorporating SRA had higher efficiency in reducing the
drying strain rate during the early period compared to that of control mixtures, but had a
lower efficiency at later stage. Furthermore, UHPC specimens incorporating SRA
exhibited about 36 and 14% reduction in the developed strain during the initial drying
period compared to that of the w/c=0.22 and 0.25 control specimens, respectively. This
strain reduction due to incorporating SRA continued to decrease with the application of
DW cycles. Moreover, submerged specimens incorporating SRA showed similar
shrinkage strain to that of the submerged control specimens with no SRA. Figure 5-10
shows the measured shrinkage strain after fixing the strain to zero at the end of the
expansion period. Therefore, it is hypothesized that SRA is washed out with migrating
water since SRA is a chemical that reduces the pore fluid surface tension but does not
Chapter 5 163
chemically combine with other hydration products (Rodden and Lange, 2004). To
validate this hypothesis, chemical oxygen demand (COD) tests were conducted on water
samples taken from the submersion tanks. The COD test is based on the fact that nearly
all organic compounds can be oxidized by strong oxidizing agents (Sawyer et al., 2003).
Specifically, the COD test measures the total amount of oxygen required to oxidize the
organic matter in a water sample, regardless of the biodegradability of the organic
substances. Hence, it is widely used for measuring water quality.
-200
-150
-100
-50
00 24 48 72 96 120 144 168
Age (hours)
Stra
in (µ
ε)
ControlSRA
Figure 5-10: Measured strain after the early expansion period under submerged
condition.
Figure 5-11 (a,b) show the COD values for the control and SRA mixtures. The
results confirm the existence of SRA in the submerging water. The cumulative SRA
percentage increased with time, which indicates continuous washout of SRA. However,
Chapter 5 164
the w/c=0.22 mixture incorporating SRA showed lower COD values compared to that of
the w/c=0.25 mixture with SRA, which indicates a reduction in the amount of washed out
SRA. Based on the COD value for organic SRA, about 17% and 22% of the SRA was
washed out from the w/c=0.22 and 0.25 mixtures by the end of the investigated period,
respectively. In the case of DW cycles, COD values decreased simultaneously with DW
cycle repetition, especially for the w/c=0.22 mixture, which showed a significant
reduction in the COD values after the first cycle, as shown in Fig. 5-11(b) This can be
attributed to the reduction and depercolation of capillary pores by hydration products as
mentioned before. In addition, the ability of SRA to migrate to the curing water decreased
as the SRA concentration in the specimens decreased with time due to SRA leaching.
0
500
1000
1500
2000
2500
3000
3500
1 3 5 7 9Number of cycles
CO
D (m
g/L)
(1.3
x10-4
oz/
gal)
0
5
10
15
20
25
30
35
Cum
ulat
ive
SRA
loss
(%)
Control-w/c=0.22SRA-w/c=0.22Control-w/c=0.25SRA-w/c=0.25Cumulative SRA loss-w/c=0.22Cumulative SRA loss-w/c=0.25
(a)
Figure 5-11: COD values for control and SRA UHPC specimens under a)
submerged condition and b) drying/wetting cycles.
Chapter 5 165
0
100
200
300
400
500
600
1 2 3 5 7 9Number of cycles
CO
D (m
g/L)
(1.3
x10-4
oz/
gal)
0
1
2
3
4
5
6
SRA
loss
(%)
w/c=0.22- COD values
w/c=0.25- SRA lossw/c=0.22- SRA lossw/c=0.25- COD values
(b)
w/c=0.22- SRA lossw/c=0.22- SRA lossw/c=0.25- SRA lossw/c=0.25- SRA lossw/c=0.25- SRA lossw/c=0.25- SRA loss
Figure 5-11 Cont’d: COD values for control and SRA UHPC specimens under a)
submerged condition and b) drying/wetting cycles.
5.4.2.3 Effect of SAP
Figure 5-12 illustrates the shrinkage strain development for w/c=0.25 UHPC specimens
incorporating SAP under different curing conditions. Regardless of the w/c, UHPC
specimens incorporating SAP showed higher mass loss during drying than that of the
control mixtures, yet showed a significantly lower total shrinkage. This can be ascribed to
the effect of the pore size distribution and developed capillary tensile stress (Wittmann,
1982). According to the Laplace law (Eq. 5-2), this capillary stress is highly affected by
the radius of the pore ( ) in which the meniscus forms. Therefore, water loss from
smaller pores will probably induce higher capillary tensile forces, leading to higher
shrinkage (Aly and Sanjayan, 2009, Collins and Sanjayan, 2000).
sr
Chapter 5 166
sc r
θγσ cos2 ⋅−=
Eq. 5-2
Where cσ is the capillary stress, γ is the surface tension of the pore solution, θ is the
contact angle between the pore solution and the solid (often assumed to be 0°), and is
the average radius of meniscus curvature.
sr
-1000
-800
-600
-400
-200
0
200
400
0 24 48 72 96 120 144 168
Age (hours)
Stra
in (µ
ε)
DryAutogenousDrying/wettingSubmerged
Submerged
Dry
Autogenous
Drying/wetting
Figure 5-12: Measured strain for w/c=0.25 UHPC specimens incorporating 0.6%
SAP under different exposure conditions.
Analysis of the incremental pore size distribution data from MIP tests showed that
SAP specimens had a proportion of finer pore size similar to that of the control
specimens as shown in Fig. 5-13. The different pore sizes were classified according to the
International Union of Pure and Applied Chemistry system (IUPAC) (IUPAC, 1972).
The calculated ( ) values were higher for SAP specimens than that for the control sr
Chapter 5 167
specimens with no SAP, as shown in Table 5-1. This is expected to result in lower
capillary stress, leading to lower shrinkage regardless the mass loss. The parameter ( )
was calculated following the same procedure reported by (Aly and Sanjayan, 2009,
Collins and Sanjayan, 2000). Furthermore, since the UHPC tested had a very low w/c
(i.e. w/c=0.22 and 0.25), the contribution of autogenous shrinkage to the total shrinkage
cannot be ignored. Therefore, a percentage of the total shrinkage reduction can be
ascribed to a reduction in the autogenous shrinkage induced by SAP in drying specimens
(Jensen and Hansen, 2002).
sr
0
20
40
60
80
100
1 2 5 9 1 2 5 9
Number of cycles
Pore
s si
ze p
erce
ntag
e (%
)
Meso Macro Voids
Control mixture SAP mixture
Figure 5-13: Pore size percentage for w/c=0.25 control and SAP UHPC specimens
under drying conditions.
Chapter 5 168
Table 5-1: Computation of ( ) for control and SAP mixtures under drying conditions and drying/wetting cycles
sr
Control SAP
Drying conditions
Number of cycles 1 5 9 1 5 9 % Moisture loss from 1.50 2.00 2.20 1.68 2.43 2.55
Volume of water dried 3.10 4.20 4.60 3.55 5.18 5.51
Total porosity (from MIP) 14.8 11.1 9.66 14.51 11.4 10.6
sr (nm) 70.0 46.0 16.00 90.00 65.0 22.0
DW cycles
Number of cycles 3 5 9 3 5 9 % Moisture loss from 1.19 1.42 1.13 1.27 1.33 1.22
Volume of water dried 2.54 3.04 2.48 2.72 2.89 2.59
Total porosity (from MIP) 12.5 10.2 9.56 13.83 9.69 8.57
sr (nm) 65.0 38.0 12.00 36.00 20.0 8.00
On the other hand, SAP specimens showed mass loss and net shrinkage under DW
cycles higher than those of the control specimens at different w/c ratios. Moreover, the
shrinkage rate was much higher than that of the control specimens. Analysis of the TGA
and MIP results for UHPC specimens incorporating SAP under DW cycles indicates that
such specimens had higher degree of hydration and lower value of the parameter ( )
compared to that of the control specimens subjected to DW cycles as shown in Fig. 5-14
and Table 5-1. SAP particles have a high tendency to absorb water and expand (Jensen
and Hansen, 2001). Hence, it is expected that SAP will absorb water during the wetting
period, leading to higher expansion as a result of SAP particles swelling. During the
drying period, SAP particles release water and occupy smaller volume (Wang et al.,
2009), thus leading to an additional shrinkage strain along with drying and autogenous
strains. Furthermore, SAP enhances the hydration process leading to a higher degree of
sr
Chapter 5 169
hydration and smaller capillary pores. The smaller the capillary pores (i.e. parameter
( )), the higher the capillary stress induced by water loss during the drying period, thus
resulting in higher shrinkage.
sr
30
35
40
45
50
55
0 1 2 3 4 5 6 7 8 9 1Number of cycles
Deg
ree
of h
ydra
tion
(%)
0
ControlSAP
Figure 5-14: Degree of hydration for w/c=0.25 control and SAP UHPC specimens
under drying/wetting cycles.
However, UHPC specimens incorporating SAP gained less mass under SM
conditions, but exhibited higher early expansion compared to that of the control
specimens. For instance, after 24 hours specimens incorporating SAP exhibited about
70% and 81% higher expansion strains compared to that of the control specimens with
w/c=0.22 and 0.25, respectively. Moreover, SAP specimens showed a significant
reduction in the developed shrinkage after the initial swelling compared to that of the
control and SRA specimens. As a result, the net strain at the end of the investigation
Chapter 5 170
period was an expansion (about 122 and 237 µε for mixtures with w/c=0.22 and 0.25,
respectively). During the early period, submerged specimens absorb water leading to
mass increase and swelling, as can be observed in Figs. 5-3 and 5-12. On the other hand,
sealed specimens showed a high autogenous shrinkage (Fig. 5-12), which is likely
attributed to self-desiccation (Ma et al., 2004). Therefore, it can be expected that the
imbibed water effectively mitigates self-desiccation during the early period rather than
the SAP itself. Once depercolation of capillary pores occurred, water stored inside SAP
particles started its role in mitigating self-desiccation. This would explain the relatively
lower shrinkage rate after the initial swelling for submerged SAP specimens compared to
that of the control specimens. Figure 5-15 captures this behaviour in which the strain
curves for the SM condition were initiated to zero at the end of the swelling. The
reduction in shrinkage strains after the early expansion was about 52% and 66 % for
mixtures incorporating SAP compared to that of the control mixtures with w/c=0.22 and
0.25, respectively.
Chapter 5 171
-250
-200
-150
-100
-50
00 24 48 72 96 120 144 168
Age (hours)
Stra
in (µ
ε)
Control-w/c=0.22SAP-w/c=0.22Control-w/c=0.25SAP-w/c=0.25
Figure 5-15: Measured strain for control and SAP UHPC specimens under SM
condition after early expansion
5.4.2.4 Synergistic Effect of SRA and SAP
To investigate a possible synergistic effect of SRA and SAP, an additional mixture with
w/c=0.25 and incorporating both 2% SRA and 0.6% SAP was tested. This mixture (noted
MSS) achieved higher efficiency in reducing shrinkage strains under different exposure
conditions. Table 5-2 shows the percentages of shrinkage strain reduction of MSS
compared to those of the other mixtures. Under SM conditions, MSS specimens showed
high expansion during early-age, similar to that of UHPC specimens incorporating SAP,
resulting in a net shrinkage strain reduction of about 70% and 71% compared to that of
the control and SRA specimens. Moreover, the measured COD for MSS specimens
indicated only 5% reduction in SRA loss compared to that of the SRA specimens. Hence,
SAP can be considered to have the dominant effect under SM conditions.
Chapter 5 172
Under DW cycles, MSS specimens showed a significant reduction in shrinkage
strains compared to that of the other mixtures as shown in Table 5-2. Moreover, COD
results indicated a significant reduction in SRA loss during DW cycles. After the first DW
cycle, the net measured COD value for the MSS mixture was about 525 mg/L, which is
similar to that of the SRA mixture. At the end of the second cycle, the net measured COD
value was about 129 mg/L compared to 372 mg/L for the SRA mixture (about 65%
reduction). On the other hand, MSS specimens showed lower mass loss and shrinkage
strains compared to that of the SAP mixture specimens. This can be explained as follows:
SAP supplies water for further hydration and consequently a denser pore structure is
developed. This, in turn leads to a smaller meniscus radius which results in higher
capillary stresses. However, the developed denser pore structure also reduces the amount
of washed out SRA, which decreases the surface tension of the pore water, leading to
lower mass loss and capillary stresses. Hence, the synergistic effect of SRA and SAP
optimizes the mutual benefits of these shrinkage mitigation methods and could overcome
their individual deficiencies in the various exposure conditions. However, further
research is needed to investigate this synergistic mechanism in order to possibly develop
a new generation of high- performance shrinkage mitigation admixtures.
Table 5-2: Percentage of reduction in shrinkage strain for each exposure condition
due to incorporating both SRA and SAP compared to reference mixtures.
% Reduction in strain due incorporating both SRA and SAP
Exposure conditions Reference
mixtures Sealed Drying Drying/Wetting Submerged
Control 50.0% 52.0% 44.0% 70.0% SAP 39.0% 40.0% 49.0% 16.0%
SRA 35.0% 42.0% 32.0% 71.0%
Chapter 5 173
5.5. CONCLUSIONS
In this study, UHPC mixtures with different w/c and with or without SRA and SAP were
tested under sealed, SM and DW cycles conditions. Based on this work, the following
conclusions can be drawn:
1) Environmental exposure conditions alter the shrinkage behaviour of UHPC with
and without shrinkage mitigation methods.
2) Initial high early swelling of submerged specimens results in very low net
shrinkage strain.
3) The washout of SRA during SM and/or DW cycles can dismiss its effectiveness in
mitigating shrinkage strains.
4) Using SAP under DW cycles can result in higher shrinkage strains of UHPC since
it stimulates a higher degree of hydration (i.e. smaller capillary pores) leading to
higher capillary tensile stress.
5) Using SAP under submerging condition is very effective in reducing the
developed shrinkage strain.
6) Using a combination of SRA and SAP creates a synergistic effect whereby the
benefits of these shrinkage mitigation methods are optimized. This allows
overcoming their individual deficiencies under different exposure conditions,
which is promising for developing a new generation of high-performance
shrinkage mitigation admixture with dual effect.
Chapter 5 174
5.6. REFERENCES
Acker, P. and Behloul, M., (2004), “Ductal ® technology: A large spectrum of properties, a wide range of applications” In: Proceedings of the International Symposium on UHPC, Kassel, Germany, pp. 11-23.
Aïtcin, P.C., (1998). High-Performance Concrete, 4th Edition, Taylor & Francis, London,
591 p. Alford, N.N. and Rahman, A.A., (1981), “An assessment of porosity and pore sizes in
hardened cement pastes,” Journal of Materials Science, Vol. 16, No. 11, pp. 3105-3114.
Aly, T. and Sanjayan, J.G., (2009), “Mechanism of early age shrinkage of concretes,”
Materials and Structures, Vol. 42, No. 4, pp. 461-468. Bentz, D.P. and Garboczi, E.J., (1991), “Percolation of phases in a three-dimensional
cement paste microstructural model,” Cement and Concrete Research, Vol. 21, No. 2-3, pp. 325-344.
Bentz, D.P., Geiker, M.R. and Hansen, K.K., (2001), “Shrinkage-reducing admixtures
and early-age desiccation in cement pastes and mortars,” Cement and Concrete Research, Vol. 31, No. 7, pp. 1075-1085.
Bentz, D.P. and Jensen, O.M., (2004), “Mitigation strategies for autogenous shrinkage
cracking,” Cement and Concrete Composites, Vol. 26, No. 6, pp. 677-685. Bentz, D.P., (2006), “Influence of shrinkage-reducing admixtures on early-age properties
of cement pastes,” Journal of Advanced Concrete Technology, Vol. 4, No. 3, pp. 423-429.
Cheyrezy, M., Maret, V. and Frouin, L., (1995), “Microstructural analysis of RPC
(Reactive Powder Concrete),” Cement and Concrete Research, Vol. 25, No. 7, pp. 1491-1500.
Chunqiu, L., Kefei, L. and Zhaoyuan, C, (2008), “Numerical analysis of moisture
influential depth in concrete during drying-wetting cycles,” Tsinghua Science and Technology, Vol. 13, No. 5, pp. 696-701.
Collins, F. and Sanjayan, J.G., (2000), “Effect of pore size distribution on drying
shrinking of alkali-activated slag concrete,” Cement and Concrete Research, Vol. 30, No. 9, pp. 1401-1406.
Cook, R.A. and Hover, K.C., (1999), “Mercury porosimetry of hardened cement pastes,” Cement and Concrete Research, Vol. 29, No. 6, pp. 933-943.
Cusson, D. and Hoogeveen, T.J., (2006), “Measuring early-age coefficient of thermal
expansion in high-performance concrete,” International RILEM Conference on
Chapter 5 175
Volume Changes of Hardening Concrete: Testing and Mitigation, Lyngby, Denmark RILEM Publications SARL, pp. 321-330.
Cusson, D. and Hoogeveen, T.J., (2007), “An experimental approach for the analysis of
early-age behaviour of high-performance concrete structures under restrained shrinkage,” Cement and Concrete Research, Vol. 37, No. 2, pp. 200-209.
Dhir, R.K. and McCarthy, M.J., (1999), “Concrete Durability and Repair Technology.”
Proceedings of International Conference : Creating with Concrete, Dyer T.D., Henderson, N.A, and Jones R.M (ed.), Thomas Telford, Scotland, U.K., 794 p.
Garas, V.Y., Kahn, L.F. and Kurtis, K.E., (2009), “Short-term tensile creep and shrinkage
of ultra-high performance concrete,” Cement and Concrete Composites, Vol. 31, No. 3, pp. 147-152.
Hannah, R.L. and Reed S.E., (1992). Strain Gage Users' Handbook, 1st Edition, Springer,
New York, 496 p. Holschemacher; K., Weiße, D. and Klotz, S., (2004), “Bond of Reinforcement in Ultra
High Strength Concrete”, Proceedings of the International Symposium on UHPC, Kassel, Germany, pp. 375-388.
Holschemacher, K. and Weiße, D., (2005), “Economic mix design of ultra high strength
concrete” Proceedings of the 7th International Symposium on Utilization of High-Strength/ High-Performance Concrete, Washington D.C., Vol. 2, pp.1133-1144.
Ichinomiya, T., Hishiki, Y., Ohno, T., Morita, Y. and Takada, K., (2005), “Experimental
study on mechanical properties of ultra-high-strength concrete with low-autogenous-shrinkage,” ACI Materials Journal, Vol. 228, pp. 1341-1352.
IUPAC, (1972). Manual of Symbols and Terminology, Appendix 2, Part 1. In: Colloid
and Surface Chemistry, Pure Applied Chemistry, Vol. 31, 578 p. Jensen, O.M. and Hansen, P.F., (2001), “Water-entrained cement-based materials: I.
Principles and theoretical background,” Cement and Concrete Research, Vol. 31, No. 4, pp. 647-654.
Jensen, O.M. and Hansen, P.F., (2002), “Water-entrained cement-based materials: II.
Experimental observations,” Cement and Concrete Research, Vol. 32, No. 6, pp. 973-978.
Kovler, K., (1996), “Why sealed concrete swells,” ACI Materials Journal, Vol. 93, No. 4, pp. 334-340.
Lea, F.M. and Hewlett, P.C., (1998). Lea's Chemistry of Cement and Concrete, 4th
Edition, John Wiley & Sons, New York, 1057 p.
Chapter 5 176
Loukili, A., Khelidj, A. and Richard, P., (1999), “Hydration kinetics, change of relative humidity, and autogenous shrinkage of ultra-high-strength concrete,” Cement and Concrete Research, Vol. 29, No. 4, pp. 577-584.
Lstiburek, J. and Carmody, J., (1994). Moisture Control Handbook: Principles and
Practices for Residential and Small Commercial Buildings, 2nd Edition, John Wiley & Sons, New York, 232 p.
Ma, J., Dehn, F., Tue, N.V., Orgass, M. and Schmidt, D., (2004), “Comparative
investigation on ultra-high performance concrete with and without coarse aggregate,” Proceedings of the International Symposium on UHPC, Kassel, Germany, pp. 206-212.
Martys, N.S. and Ferraris, C.F., (1997), “Capillary transport in mortars and concrete,”
Cement and Concrete Research, Vol. 27, No. 5, pp. 747-760. Neville, A.M., (1996). Properties of Concrete, 4th Edition, John Wiley &Sons, New York,
797 p. Ribeiro, A.B., Gonçalves, A. and Carrajola, A., (2006), “Effect of shrinkage reducing
admixtures on the pore structure properties of mortars,” Materials and Structures, Vol. 39, No. 2, pp. 179-187.
RILEM Report 25, (2003), “Early-age Cracking in Cementitious Systems,” In: RILEM
Technical Committee 181-EAS: Early-age Shrinkage Induced Stresses and Cracking in Cementitious Systems, A. Bentur (ed.), 335 p.
Rodden, R.A. and Lange, D.A., (2004), “Feasibility of shrinkage reducing admixtures for
concrete runway pavements,” Technical note No. 4, The Center of Excellence for Airport Technology (CEAT), 6 p.
Sawyer, C.N., McCarty, P.L. and Parkin, G.F., (2003). Chemistry for Environmental
Engineering and Science, 5th Edition, McGraw-Hill Professional, New York, 752 p.
Schmidt, M. and Fehling, E., (2005), “Ultra-high-performance concrete: research,
development and application in Europe” Proceedings of the 7th International Symposium on Utilization of High-Strength/High-Performance Concrete, Washington D.C., Vol. 1, pp. 51-77.
Sule, M. and Van Breugel, K., (2001), “Cracking behavior of reinforced concrete
subjected to early-age shrinkage,” Materials and Structures, Vol. 34, No. 5, pp. 284-292.
Chapter 5 177
Tazawa, E., (1999), “Autogenous Shrinkage of Concrete,” Proceedings of the International Workshop organized by the Japan Concrete Institute, Hiroshima, Japan, Routledge, Taylor & Francis, New York, 411 p.
Teichmann, T. and Schmidt, M., (2004), “Influence of packing density of fine particles
on structure, strength and durability,” Proceedings of the International Symposium on UHPC, Kassel, Germany, pp. 227-237.
Turcry, P., Loukili, A., Barcelo, L. and Casabonne, J.M., (2002), “Can the maturity
concept be used to separate the autogenous shrinkage and thermal deformation of a cement paste at early age?,” Cement and Concrete Research, Vol. 32, No. 9, pp. 1443-1450.
Wang, F., Zhou, Y., Peng, B., Liu, Z. and Hu, S., (2009), “Autogenous Shrinkage of
Concrete with Super-Absorbent Polymer,” ACI Materials Journal, Vol. 106, No. 2, pp. 123-127.
Weiss, J., Lura, P., Rajabipour, F. and Sant, G., (2008), “Performance of shrinkage-
reducing admixtures at different humidities and at early ages,” ACI Materials Journal, Vol. 105, No. 5, pp. 478-486.
Wittmann, F.H., (1982), “Creep and shrinkage mechanisms,” In: Creep and Shrinkage in
Concrete Structures, Bazant, Z.P. and Wittmann, F.H. (ed.), Wiley, Chichester, pp. 129-161.
Yazici, H., (2007), “The effect of curing conditions on compressive strength of ultra high
strength concrete with high volume mineral admixtures,” Building and Environment, Vol. 42, No. 5, pp. 2083-2089.
Chapter 6 178
CHAPTER SIX
SELF-RESTRAINING SHRINKAGE CONCRETE: MECHANISMS AND EVIDENCE *
In the previous chapters, the effect of environmental conditions (i.e. field-like conditions)
and active shrinkage mitigation techniques on the early-age shrinkage behaviour of
UHPC was investigated. It was proven that the addition of shrinkage reducing admixture
and/or superabsorbent polymer can significantly reduce shrinkage under different
exposure conditions. However, a number of disadvantages including, reduction in
mechanical properties, retarding hydration process, higher water loss, higher porosity,
washing out effect and cost, are still limiting these materials applications. Further, these
shrinkage mitigation techniques do not provide a reduction in the environmental impact
of concrete production. In this chapter, a passive shrinkage mitigation technique is
introduced as a solution that provides adequate benefits for both facets.
6.1. INTRODUCTION
Once water comes in contact with cement, exothermic chemical reactions (so-called
hydration reactions) initiate, leading to the formation of hydration products (Neville,
1996). These hydration products connect un-hydrated cement particles and other concrete
constituents together to form the internal microstructure of concrete (Struble et al., 1980).
As hydration reactions progress with time, the concrete microstructure gains rigidity and
increased ability to resist different stresses (Roy and Idorn, 1993).
*Some parts of this chapter (Part 1) were accepted by the ACI Materials Journal. Other parts (Part 2) have been submitted for review to Cement and Concrete Research.
Chapter 6 179
On the other hand, the volume occupied by hydration products will be less than
that occupied by its reactants before hydration, leading to a volume contraction. This
internal deformation is known as chemical shrinkage (Le Chatelier contractions) (Mehta
and Monteiro,2006). Under sealed conditions, the progress of hydration reactions and
associated chemical shrinkage will result in an external bulk deformation, which is
referred to as autogenous deformation (Kovler and Zhutovsky, 2006, Hansen, 2011).
Hence, chemical shrinkage can be considered as the main driving force behind
autogenous deformation (Barcelo et al., 2001). Such an early-age shrinkage of concrete is
a complex physico-chemical phenomenon, which is related to the hydration reactions and
progressive hardening of the concrete skeleton.
This chapter describes a new concept for reducing early-age shrinkage in
hydrating cement-based materials through the addition of partially hydrated cementitious
materials (PHCM) as a concrete admixture. During the mixing of concrete, the added
PHCM will form restraining clusters, which essentially consist of a hydrated cement
paste structure and micro-crystals of hydration products. Conceptually, this is similar to
the role played by un-hydrated cement particles and aggregates, which impart a passive
internal restraint within concrete (Bentz and Jensen, 2004). The PHCM technique creates
deformation resistance as early as the cementitious materials are mixed through
developing a connected strength-gaining skeleton at the fresh stage. An explanation of
the mechanism and rationale of the PHCM technique is given below; validity and
benefits are also discussed below.
Chapter 6 180
6.2. EARLY-AGE PHASES AND AUTOGENOUS SHRINKAGE IN
CEMENTITIOUS MATERIALS
Early-age shrinkage is defined as a volume change occurring immediately after placing
concrete up to the age of about 24 hours (Holt, 2001). During this time the structural
development in concrete can be categorized into three stages (Esping, 2007) (i) liquid
stage when concrete is plastic, workable, and having a visco-elastic behaviour, (ii)
skeleton formation stage (semi-plastic) when concrete undergoes early stiffening and a
transition from fluidity to rigidity through the progressive forming a self-supporting
skeleton which usually begins after the initial setting), and (iii) hardened (rigid) stage
which starts at the point of final setting when the concrete starts developing mechanical
strength as a result of continuing hydration reactions.
At the liquid stage, when no significant skeleton that restrains deformation exists,
autogenous shrinkage is considered to be equal to chemical shrinkage (Barcelo et al.,
2001, Holt, 2001). Once concrete enters the skeleton formation stage, a self-supporting
skeleton starts to form; hence, autogenous shrinkage will diverge from chemical
shrinkage since the rigidity of the paste restrains volume changes (Mehta and
Monteiro,2006). At this stage, the rate of autogenous shrinkage will decrease to about
(1/10) of the rate of chemical shrinkage (Hammer, 1999a). In conclusion, autogenous
shrinkage initially originates from the chemical shrinkage, once the concrete set the
development of a contracting capillary pore pressure will control the subsequent
deformations (Hansen, 2011).
Chapter 6 181
Autogenous and chemical shrinkage are influenced by several factors. However,
these factors can have similar or conflicting effects on autogenous and chemical
shrinkage. Indeed, factors that increase chemical shrinkage may reduce autogenous
shrinkage, for instance when they reduce the setting time and/or generate a more rigid
restraining structure (Esping, 2007).
6.3. EARLY-AGE SHRINKAGE MITIGATION TECHNIQUES
Early-age shrinkage can be more likely to cause cracking than long-term shrinkage since
it develops more rapidly and occurs when the cement-based material still has low
mechanical properties (Holt and Leivo, 2004). When formed, cracks may reduce strength,
jeopardize durability, cause loss of pre-stress, etc. Therefore, different strategies for
mitigating early-age shrinkage cracking have been proposed.
Understanding the relationships amongst forces existing in a cement paste during various
early-age phases, their sources and how they affect the overall behaviour is key to
tailoring effective shrinkage mitigation strategies. According to the Laplace equation
(Eq. 6-1), at the maximum capillary pore radius filled with water, water is under tensile
stresses:
rPca
)cos(2 θσ ⋅⋅−= Eq. 6-1
Where is the tensile stress in the pore solution, caP σ is the surface tension of the pore
solution, θ is the wetting angle of the solid with pore solution, and r is the radius of the
meniscus curvature. The tensile stresses developed in the capillary water must be
Chapter 6 182
balanced by compression stresses on the surrounding solid (Kovler and Zhutovsky,
2006). During the liquid phase, the induced capillary stresses are greater than the bonding
strength among cement paste particles; hence, the particles will be rearranged in the
water. At that time, autogenous shrinkage is considered to be equal to chemical shrinkage
(Barcelo et al., 2001, Tazawa, 1999). Once a solid path within the material is developed,
the material rigidity gradually increases due to increasing connectivity of the solid
particles; the bond strength among particles becomes greater than capillary stresses. This
moment is known as the mineral percolation threshold (Barcelo et al., 2001).
Consequently, the developed skeleton starts restraining particle displacements;
however, it does not yet have sufficient strength to resist compressive stresses, which
results in a volume decrease (i.e. bulk shrinkage). Meanwhile, water will move within the
paste without particle displacements. Hence, autogenous shrinkage starts to diverge from
chemical shrinkage (Justens et al., 1996, Hammer, 1999b). As hydration reactions
progress, the paste skeleton gains increased strength to resist further three-dimensional
deformation and air voids develop within capillary pores, leading to self-desiccation
(Barcelo et al., 2001). Once the breakthrough pressure has been reached, the water
meniscus cannot find a stable position due to discontinuity of the water system. Hence, a
collapse of capillary pressure takes place (Wittmann, 1976). Therefore, early-age
shrinkage of cementitious materials is mainly related to the development of capillary
stresses, internal self-desiccation, and the strength of the restraining skeleton. Figure 6-1
illustrates a correlation between cement paste structure development and volume change
during different early-age stages.
Chapter 6 183
Figure 6-1: Structure development and volume change during different early-age
stages.
Considering the various mechanisms discussed above, several concrete shrinkage
mitigation techniques have been proposed. The first mitigation strategy deals with the
surface tension of the pore solution. The latter is indeed a critical parameter for
autogenous shrinkage as it directly affects capillary stress based on Laplace's equation
(Eq. 6-1). The most commonly used material in this category is shrinkage-reducing
admixtures (SRA). SRA decrease the surface tension of the pore solution, hence, it is
considered as a direct method for reducing autogenous shrinkage (Tazawa and
Miyazawaa, 1995a). However, possible disadvantages of SRA include a reduction of its
effectiveness over time due to absorption of SRA by hydration products, destabilization
of the entrained air void system (Schemmel et al., 1999), delaying the setting time and
retardation of hydration reactions, thus leading to reduced mechanical strength (Bentz,
2006). Moreover, the surface tension is also influenced by temperature, which may result
in dosage variations to achieve optimum behaviour (Jensen and Hansen, 1999). In
addition, washing out effect of SRA can significantly reduce it shrinkage mitigation
efficiency as illustrated in chapter 5.
Chapter 6 184
The second shrinkage mitigation strategy deals with both the radius of the water
meniscus of the largest water-filled pore within the microstructure and the availability of
water for hydration. These two parameters have a significant effect on capillary stresses
and the internal self-desiccation process. In this strategy, internal curing techniques have
been proposed, including the use of saturated lightweight aggregate particles with coarser
pores (i.e. larger radius) (Bentur et al., 2001). This results in the development of lower
capillary stress since the formation of empty pores due to chemical shrinkage first takes
place in such coarser pores (in the aggregate) and does not involve relatively finer pores
in the cement paste (Henkensiefken et al., 2010). In addition, the water contained within
the saturated aggregate pores provides internal curing for the hydrating cement paste.
Nevertheless, difficulties in controlling the consistency of the rheological properties of
concrete incorporating such aggregates, along with potential reductions in its mechanical
strength and elastic modulus constitute limitations of this technique. More recently, based
on a similar principle to that of saturated lightweight aggregate particles, superabsorbent
polymer particles (SAP) have been proposed as a shrinkage mitigation admixture for
concrete (Jensen and Hansen, 2001). SAP is a more direct technique which produces a
better controlled microstructure since it uniformly distributes the curing water.
While significant research has investigated the above shrinkage mitigation
strategies, to the best of the authors’ knowledge, research dedicated to exploring the
effect of a restraining skeleton on shrinkage is rather scarce. Therefore, the present study
focuses on the restraining skeleton concept and pioneers a self-restraining shrinkage
system to reduce early-age shrinkage of concrete through the addition of partially
hydrated cementitious material.
Chapter 6 185
6.4. SHRINKAGE RESTRAINING SYSTEMS
The concept of a shrinkage restraining system itself is not new. Generally, un-hydrated
cement particles and aggregates constitute an internal restraining system (Bentz and
Jensen, 2004), which resists autogenous deformation and reduces physical shrinkage
(Tazawa and Miyazawaa, 1995b). However, using a higher aggregate content for
reducing autogenous shrinkage in modern concretes such as high performance concrete
or self-consolidating concrete is not practical. This is because increasing the aggregate
content leads to reducing the paste volume, which is detrimental to workability in low
w/c systems (Kovler and Zhutovsky, 2006) or where aggregate friction causes blockage
of flow.
One of the key parameters in early-age autogenous shrinkage is the time of
hardening (Holt, 2001). This corresponds to the time when the cementitious material has
sufficient strength to resist shrinkage stresses. The hydrated cement paste in fresh
concrete forms a self-restraining system that reduces the absolute volume contraction
(Aitcin, 1999). Therefore, accelerating agents and heat curing which cause an earlier
setting and structuring of a stiff skeleton will shorten the period during which early-age
autogenous shrinkage occurs (Holt, 2001, Kronlöf et al., 1995). However, such
accelerating methods also accelerate the rate of autogenous shrinkage along with
increasing mechanical strength.
At the micro-level, calcium hydroxide (CH) crystals act as a passive restraint and
reduce the measured physical shrinkage (Jensen and Hansen, 1996). The hypothesis that
CH micro-crystals provide restraint was proposed by (Bentz and Jensen, 2004) based on
results reported in (Carde and Francois, 1997, Powers, 1962) which indicate a significant
Chapter 6 186
effect of leaching of CH crystals on deformation properties. The role of CH micro-
crystals is schematically illustrated in Fig. 6-2.
Figure 6-2: Micro-crystals role in restraining deformation.
6.5. CAN PHCM ADDITION REDUCE CONCRETE SHRINKAGE
The ability of PHCM addition to reduce early-age concrete shrinkage in concrete is
affirmative at least for the main reasons below.
First, a cement paste matrix can be divided into two phases, the liquid phase and
solid phase (i.e. un-hydrated particles and hydration products). As discussed earlier,
emptying of capillary pores in a cementitious material via drying or chemical reactions
subjects capillary water to tensile stresses and the solid paste to compressive stresses,
which leads to volume reductions (Kovler and Zhutovsky, 2006). The compressive
pressure that induces such stresses on the solid paste represents effective pressure
transmitted through points of contact between the solid phase particles (Fosså, 2001).
This effective pressure can be calculated from Eq. 6-2:
Chapter 6 187
⎟⎠⎞
⎜⎝⎛ −
−=′A
AAupp c Eq. 6-2
Where p' is the effective pressure, p is the total pressure, u is the pore water pressure, A is
the total area, and Ac is the particle contact area. Increasing contact points between grains
increases the resistance to apparent volume changes (Barcelo et al., 2005). During the
liquid stage, the bonds and contact areas amongst particles are very small. Hence, the
concrete does not resist significant stresses or deformations (Fosså, 2001). As hydration
reactions proceed, the contact areas and bond between particles increase, resulting in
increased ability to resist deformation. The higher the contact area, the higher is the
resistance. Adding PHCM increases the contact areas acting as a load bearing structure
for induced compressive stresses. This reduces volume reduction and overall
deformation.
Second, the added PHCM provides a sufficient initial amount of hydration
product crystals (i.e. calcium silicate hydrate (CSH) and CH) and induces a hydrated
cement paste structure within the original otherwise fresh mixture at the onset of the
hydration process. In other words, the PHCM increases the passive internal restraint
within the fresh cementitious matrix, which is analogous to the role of un-hydrated
cement and aggregate particles. This can be illustrated as follows: in concrete, the
hydrated cement paste constitutes a continuous matrix phase; shrinkage can be reduced
due to the presence of discrete restraining particles (i.e. un-hydrated cement and
aggregates). The volume of the shrinkable paste can be expressed as follows (Hansen,
1987):
Chapter 6 188
)( ucacpc vvvv +−= Eq. 6-3
Where is the volume of paste in concrete, is the volume of concrete, is the
volume of aggregate, and is the volume of un-hydrated cement. Aggregates and un-
hydrated cement particles are assumed to be the sole shrinkage restraining agents.
However, concrete contains partially hydrated cement. Thus, the volume of un-hydrated
cement should be expressed as follows (Hansen, 1987):
pcv cv av
unv
)1( α−= coun vv Eq. 6-4
Where is the volume of cement added originally and cov α is the relative degree of
hydration, which varies between 0 and 1. Substituting Eq. 6-4 into Eq. 6-3, one obtains
the following equation:
[ ])1( α−+−= coacpc vvvv Eq. 6-5
Therefore, the relative shrinkage restraining volume ( ) can be defined as follows
(Hansen, 1987):
dV
c
coad v
vvV
))1(( α−+= Eq. 6-6
For mixtures incorporating PHCM, the added PHCM can be considered as an additional
internal shrinkage restraining material, the volume of shrinkage paste can be modified to
the following:
Chapter 6 189
[ ] [ ] ηαααηα ⋅+−−+−+−⋅−+−= 1212 )1)(1(()1())1(( cocoaccoacpc vvvvvvvv
Part 1 Part 2
Eq. 6-7
Where Part 1 represents the volume of paste in the mixture batch before adding PHCM,
while Part 2 represents the volume of paste in the added PHCM; 1α is the relative
degree of hydration of the paste in the added PHCM, η is the percentage of the added
PHCM; 2α is the relative degree of hydration of the paste without PHCM addition. By
rearranging Eq. 6-7, it can be reduced to the follow:
[ ] [ ]))1()1( 212212 ααηαααηα ⋅⋅⋅+−+−=⋅⋅+−+−= cocoaccoacpc vvvvvvvv Eq. 6-8
Hence, the relative restraining volume ( ) can be reformulated to include the effect of
the PHCM as follow:
dV
c
co
c
coad v
vv
vvV 212 )1( ααηα ⋅⋅⋅
+−+
= Eq. 6-9
Comparing the Eq. 6-6 and Eq. 6-9, a new term (i.e.cv
21cov α⋅α⋅η⋅ ), representing the
relative shrinkage resistance induced by the addition of PHCM, is added. Therefore,
mixtures incorporating PHCM can possess a higher relative deformation restraining
volume compared to that in normal mixtures. According to (Hansen, 1987), the
volumetric shrinkage of concrete ( cvv )(∆ can be expressed as follows:
)1( dp
pc
c
VKP
vv
−=⎟⎠⎞
⎜⎝⎛ ∆
Eq. 6-10
Where is the total hydrostatic shrinkage stress within the paste in concrete, and
is the bulk modulus of the paste. It can be deduced from Eq. 6-10 that the volumetric
pcP pK
Chapter 6 190
shrinkage of concrete decreases as V increases. Hence, it is expected that mixtures
incorporating PHCM can exhibit lower shrinkage compared to that in similar mixtures
without PHCM.
d
Third, previous work (Lobo et al., 1995) indicated that reused concrete can induce
acceleration effect. Hence, this acceleration will increase the very early-age strength
(Kronlöf et al., 1995) and structure a strong and stiff skeleton frame which withstand
forces and limit shrinkage within fresh concrete (higher passive internal restraining
system). This is consistent with Eq. 6-10 which shows that the volumetric shrinkage of
concrete is indirectly proportional to the bulk modulus of the paste. Since PHCM addition
accelerates the hydration process, resulting in a higher paste bulk modulus, a reduction in
volumetric shrinkage can be expected. The main difference between PHCM addition and
conventional acceleration techniques is that PHCM accelerates hydration reactions along
with inducing a pre-existing passive internal restraint system, which in turn reduces the
amount of deformation developed.
6.6. SOURCES OF PHCM
It is estimated that an annual average of 17.5 million cubic meters of ready mixed
concrete produced in the USA is returned to concrete batch plants (Obla et al., 2007). The
disposal of this returned concrete has serious environmental and economical implications.
However, such a returned material can be used as a source of PHCM after some
treatment. This treatment may include the use of stabilizing admixtures (Lobo et al.,
1995) to control the degree of hydration of the PHCM before adding to a newly mixed
concrete.
Chapter 6 191
6.7. PARTIALLY HYDRATED CEMENTITIOUS MATERIALS
METHODOLOGY
The main idea behind the PHCM technique is to provide a sufficient initial amount of
calcium silicate hydrate (CSH) and calcium hydroxide (CH) through adding PHCM to the
original mixture at the onset of the hydration process. This can be achieved through
mixing the concrete batch in two stages. In the first, a portion of the mixture batch is
mixed and stored under adequate curing conditions to avoid altering the quality of the
hydrated material due to drying. The second step includes conventional mixing of the
remaining portion of the batch with the pre-mixed portion from the first stage. During the
elapsed time between the two mixing steps, hydration reactions progress and hydration
products are formed within the mixed portion before remixing. The longer the elapsed
time and/or the greater the portion mixed in the first step, the higher is the hydration
product formation, which enhances the efficiency of the added PHCM as passive internal
restraint system.
6.8. EXPERIMENTAL PROCEDURE
This experimental program aims to gain an understanding of the mechanisms involved in
using PHCM. This is of primary importance in order to enable the production of self-
restraining concrete. In this study, monitoring of the hydration process and
characterization of the cement paste microstructure has been carried out on concrete
mixtures incorporating PHCM. The experimental includes two parts. Part I was devoted
to validating the addition of PHCM as shrinkage mitigation technique and to evaluate its
effect on other early-age properties. The second part focuses on explaining the
mechanism of the PHCM technique.
Chapter 6 192
6.8.1. Materials and Mixture Proportions
The materials used in this chapter were similar to that used in Chapter 3 (refer to Section
3.4.1). The chemical and physical properties of the used binders have been given in
Chapter 3 (Table 3-1). In addition, Fig. 6-3 shows the NMR spectra for the anhydrous
cement and SF used in this study. The peak of the anhydrous cement (Q0) and SF (Q4)
occur at -71 and -110 ppm, respectively. Three PHCM proportions, namely, 25, 33 and
50% of the batch weight were used to investigate the effect of the PHCM portion on the
early-age behaviour. The elapsed time between the two mixing steps was taken equal to 6
hrs according to previous hydration kinetics studies (Odler and Dörr, 1979). The selected
composition and proportions for the control and PHCM mixtures, and the characteristics
of the tested mixtures are shown in Tables 6-1 and 6-2, respectively.
Table 6-1: Composition of control and PHCM mixtures.
Mixture (mass/ cement mass) Control PHCM 25% PHCM 33% PHCM 50% Mixing Stage
Material 1 2 1 2 1 2 1 2 Cement ‐‐‐‐ 1.0000 0.2500 0.7500 0.3300 0.6700 0.5000 0.5000Silica fume ‐‐‐‐ 0.3000 0.0750 0.2250 0.0990 0.2010 0.1500 0.1500Quartz powder ‐‐‐‐ 0.4300 0.1075 0.3225 0.1419 0.2881 0.2150 0.2150Quartz sand ‐‐‐‐ 1.5300 0.3825 1.1475 0.5049 1.0251 0.7650 0.7650Water ‐‐‐‐ 0.2500 0.0625 0.1875 0.0825 0.1675 0.1250 0.1250HRWRA ‐‐‐‐ 0.0300 0.0075 0.0225 0.0099 0.0201 0.0150 0.0150
Table 6-2: Tested mixtures
Mixture PHCM
(% of the batch weight) C 0.0 H1 25.0 H2 33.0 H3 50.0
Chapter 6 193
(a) (b)
Figure 6-3: NMR spectra for anhydrous a) Portland cement, and b) Silica fume.
6.8.2. Preparation of Test Specimens and Testing Procedures
The experimental methods used in this chapter are the cubic compressive strength test,
Semi-adiabatic calorimetry, TGA, Differential scanning calorimetry (DSC),
measurements of total free shrinkage, setting time, the corrugated tube protocol for
measuring autogenous shrinkage, and solid-state 29Si MAS-NMR. The cubic compressive
strength test, Semi-adiabatic calorimetry, TGA, and measurements of total free shrinkage
method were previously explained in chapter 3. The setting time, Differential scanning
calorimetry (DSC), the corrugated tube protocol for measuring autogenous shrinkage, and
solid-state 29Si MAS-NMR are explained in the following sections.
The time of setting was measured on three replicate paste specimens using a Vicat
needle according to ASTM C191 (Standard Test Method for Time of Setting of
Hydraulic Cement by Vicat Needle).
The decomposition of different hydration phases (e.g. calcium hydroxide (CH))
can be calculated from DTG results. However, DSC tests were performed as it requires
much less time, allowing testing every 2 hrs. The UHPC samples, around 30 to 60 mg,
Chapter 6 194
were heated in a helium atmosphere at a constant rate of 10°C per minute up to 550°C.
The DSC data analysis was done using TA Instruments thermal analysis software. The
CH content was equivalent to the area (enthalpy) under the respective endothermic peaks.
The endothermic peak for CH was observed at around 440°C. The area under the curve
was related to the quantity of the material in the sample using the regression equation
obtained from the calibration graphs for the used TA Instruments machine. At least two
DSC analyses were performed on identical specimens for each mixture at each age to
ensure repeatability of results.
Autogenous shrinkage was assessed using a special measuring technique
developed by (Jensen and Hansen, 1995), where the concrete is encapsulated in thin,
corrugated polyethylene moulds. The dilatometer was equipped with automatic data-
logging and high accuracy electronic linear displacement transducers (0.001mm).
Moreover, the dilatometer was fabricated with a built-in temperature-controlled light
paraffin oil bath at 20 ± 1°C to ensure isothermal conditions during measurements, where
three replicate specimens were evaluated concurrently while being submerged in the light
paraffin oil bath (see Fig. 6-4). Autogenous deformation measurements were zeroed at
the final setting times (Bentz and Peltz, 2008).
High-resolution solid-state 29Si MAS-NMR experiments were performed in the
Biomolecular NMR facility at the University of Western Ontario using a Varian Infinity
Plus 400 spectrometer operating at 79.4 MHz with a triple-resonance Varian T3 HXY
MAS probe having 7.5 mm ZrO2 rotors rotating at 6.0 kHz. The 29Si NMR spectra were
obtained using a single-pulse experiment with a 2.5 µs pulse roughly corresponding to a
40-degree tip-angle (π/2-pulse = 5.5 µs). The chemical shifts were all relative to
Chapter 6 195
tetramethylsilane (TMS). The free induction decays (FIDs) were apodized with 100 Hz
line-broadening and zero-filled two times before Fourier transform. The experimental
error in peak position values was estimated as ±0.1 ppm. At testing age, NMR samples
hydration were stopped using freeze-drying method.
Section A-A
(a)
(b)
Figure 6-4: The corrugated tube protocol a) test setup, and b) measuring unit.
Chapter 6 196
6.9. RESULTS AND DISCUSSION
6.9.1. Setting Time and Compressive Strength
Regardless of the added amount of PHCM, the PHCM technique significantly enhanced
the development of early-age compressive strength. All PHCM mixtures consistently
produced higher early-age compressive strengths compared to that of the control mixture
(C) cured at 10 and 20°C, as shown in Fig. 6-5(a,b). The increase in compressive
strength compared to that of the C was directly proportional to the added amount of
PHCM. For instance, the 24 hrs compressive strength at 10 and 20°C increased by 70%
and 30% for H1, 85% and 40% for H2 and by 140% and 45% for H3 of their respective
C values. At 10°C, higher increase in compressive strength with PHCM addition was
achieved compared to that at 20°C. This is likely due to the influence of the added PHCM
on accelerating hydration reactions. Hence, the slow rate of hydration and strength
development at the low temperature of 10°C were compensated for, leading to more
hydration products and stronger microstructure (ACI committee 701, 2001).
Chapter 6 197
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Com
pres
sive s
treng
th (M
Pa) C
H1H2H3
T=10°C(a)
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Com
pres
sive s
treng
th (M
Pa) C
H1H2H3
(b)T = 20°C
Figure 6-5: Early-age compressive strength of UHPC mixtures incorporating
PHCM cured at a) 10°C and b) 20°C. [Maximum COV (10°C, 20°C): C (0.9%,
2.0%), H1 (1.9%, 4.8%), H2 (2.3%, 5.6%), H3 (2.7%, 2.2%)].
Chapter 6 198
Compared with the control mixture setting time, the PHCM technique also
reduced the setting time significantly as shown in Table 6-3. These results demonstrate
that the PHCM technique can be considered as an effective setting and hardening
accelerating method.
Table 6-3: Initial and final setting times for different mixtures.
Temperature10°C 20°C
Setting time (min)Mixture Initial Final Initial Final
C 510 715 460 520 H1 450 495 340 420 H2 405 465 220 270 H3 245 280 170 210
The improvement in early-age compressive strength and reduction in setting time
induced by PHCM addition can not be attributed to addition of older material. For
instance, mixture H3 after 6 hours curing at 10°C, 50% of the mixed cementitious
material is 6 hours older. Hence, 50% of the cementitious material has an age of 12
hours, while the other 50% has an age of 6 hours. Summing up 50% of the C
compressive strength at age 6 hours (0.0 MPa) and 12 hours (1.68 MPa) results in a
compressive strength about (0.84 MPa) which is much lower than that of H3 at 6 hours
(3.90 MPa). On the other hand, setting time of normal paste starts to set at about 510 min
at 10°C, hence, adding 50% of 6-hour older paste material should theoretically reduce the
setting to about 150 min. However, the setting time was only reduced to about 245 min
(which represents about 52% reduction with respect C value). The longer setting time
achieved by H3 than the theoretical one (150 min) can be attributed to the effect of
Chapter 6 199
remixing on the developed microstructure. Remixing break down the formed connection
between hydration products (Lea, 1988), hence, the H3 paste will initially have a lower
stiffness than that of paste mixed 6 hours earlier.
Figure 6-6 shows the compressive strength results for different mixtures at age 28
days. Generally, all mixtures exhibited comparable 28 days compressive strength
compared to that of the control mixture regardless the curing temperature. For instance,
the difference in 28 days compressive strength between control mixture and PHCM
mixtures at 10°C and 20°C, ranged from -1% to +6.5% and from -4% to +0.85% with
respect to the control mixture values, respectively. This slight variation in the 28 days
compressive strength can be attributed to the fact that at low w/c ratio, hydration progress
is mainly controlled by the availability of sufficient space for hydration products to form
(Lea, 1988).
0
30
60
90
120
150
180
10 20Curing Temperature (°C)
Com
pres
sive s
treng
th (M
Pa)
CH1H2H3
Figure 6-6: 28 days compressive strength of UHPC mixtures incorporating PHCM
cured at 10°C and 20°C.
Chapter 6 200
6.9.2. Degree of Hydration
Early-age strength development of concrete mixtures highly relies on the degree of
hydration achieved (Xiao and Li, 2008). The correlation between the compressive
strength and degree of hydration (represented by the amount of BW) is plotted in Fig. 6-
7. It can be observed that the relationship exhibits a linear trend for PHCM mixtures
(with R2 = 0.97). This linear relationship between compressive strength and degree of
hydration indicates a proportional relationship between the added dosages of PHCM and
the development of hydration and strength.
R2 = 0.97
0
10
20
30
40
50
0 0.02 0.04 0.06 0.08 0.1Amount of BW (mg/mg binder)
Com
pres
sive
str
engt
h (M
Pa)
H1
H2
H3
Li
Figure 6-7: Relation between degree of hydration (amount of BW) and compressive
strength development for the PHCM mixtures.
6.9.3. Heat of Hydration
A key characteristic of cementitious materials is the heat generated due to the exothermic
hydration reactions of cement. This heat is translated into a temperature increase from
Chapter 6 201
which the heat quantity developed can be evaluated (Chikh et al., 2008) in the present
study. Test results were presented graphically in terms of semi-adiabatic temperature
evolution versus time.
Mixtures incorporating PHCM had similar temperature evolution curves that
differed from that of the control mixture, as shown in Fig. 6-8, indicating a variation in
the hydration kinetics. Initially, the temperature for PHCM mixtures rised rapidly after
casting the specimen in the semi-adiabatic cell (20 min from water addition) and did not
exhibit an induction period (period in which the rate of hydration reactions slows down
significantly (Ramachandran et al., 2002)). Temperature rised up until reaching a peak of
about 42°C after about 7.5 hrs for H3, 38°C and 39°C after about 8 hrs for H1 and H2,
respectively. On the other hand, C hydration exhibited induction and acceleration periods
as shown in Fig. 6-8. The acceleration period for C was initiated at about 6 hrs later than
that of PHCM mixtures and had a temperature peak of about 36°C at around 12 hrs from
water addition. Hence, the PHCM technique effectively diminished the induction period,
leading to a continuous progress of hydration reactions and consequently shorter setting
time and higher early-age compressive strength as discussed earlier.
Chapter 6 202
20
25
30
35
40
45
0 6 12 18 24 30 36 42 48Age (hours)
Tem
pera
ture
(°C
)
CH1H2H3
(a)
Figure 6-8: Heat of hydration for UHPC mixtures incorporating PHCM.
6.9.4. Shrinkage
Shrinkage results for mixtures incorporating different contents of PHCM are shown in
Fig. 6-9. PHCM mixtures showed lower shrinkage and mass loss compared to that of the
control mixture (Figs. 9 and 10). This can be attributed to the fact that the measured
shrinkage includes drying and autogenous shrinkage (thermal deformation can be ignored
due to the small cross-section of the samples (Baroghel-Bouny et al., 2006)). Autogenous
shrinkage is strongly related to hydration reactions in which water is consumed, leading
to internal self-desiccation (without external water loss) (Power and Brownyard, 1947).
For low w/c mixtures, the amount of autogenous shrinkage can be comparable to that of
drying shrinkage (Tazawa and Miyazawa, 1999). Generally, accelerating the rate of
hydration increases the internal self-desiccation, and consequently increases the
autogenous shrinkage contribution to the measured shrinkage.
Chapter 6 203
-600
-500
-400
-300
-200
-100
00 1 2 3 4 5 6
Age (days)
Tota
l Shr
inka
ge (
µε)
7
CH1H2H3
Figure 6-9: Total shrinkage for PHCM mixtures
[Maximum COV: C (4.5%), H1 (3.1%), H2 (1.5%), H3 (4.6%)].
-1.0
-0.8
-0.6
-0.4
-0.2
0.00 1 2 3 4 5 6 7
Age (days)
Mas
s lo
ss (%
)
CH1H2H3
Figure 6-10: Mass loss for PHCM mixtures.
Chapter 6 204
However, the added PHCM will act as a passive internal restraint, thus reducing
the amount of deformation developed along with accelerating the hydration reactions. At
the micro-level, the added PHCM provides CH crystals that can act as a passive restraint
and reduce the measured physical shrinkage (Jensen and Hansen, 1996). This explanation
is consistent with autogenous shrinkage results shown in Fig. 6-11, and with DSC results
presented below. Increasing the amount of added PHCM increased the passive internal
restrain and consequently reduced the amount of autogenous shrinkage.
-750
-600
-450
-300
-150
00 12 24 36 48 60 72 84 9
Age (hours)
Aut
ogen
ous
shrin
kage
(µε)
6
CH1H2H3
Figure 6-11: Autogenous shrinkage for PHCM mixtures.
Furthermore, the evaporable water content in PHCM mixtures is probably less
than that in the control mixture at the onset of drying. This can be explained as follows:
all mixtures initially have the same amount of mixing water, however, the added mixing
water before casting the specimens for PHCM mixtures (i.e. mixing water at the second
Chapter 6 205
stage) is less than that of the ordinary mixture. For instance, for the mixture H3, half of
the mixing water is added in the first mixing stage, while the other half is added at the
second stage before casting the shrinkage specimens. Hence, PHCM mixtures possess
lower evaporable water and consequently exhibited lower mass loss as shown in Fig. 6-
10.
6.10. MECHANISM OF PHCM ACTION
The results above demonstrate that the PHCM technique does accelerate the setting and
hardening processes. Consequently, it accelerates the rigid skeleton formation and
development of restraining system within the microstructure. The subsequent part of this
study focuses on explaining PHCM mechanisms. The H2 mixture containing a moderate
dosage of PHCM will be used in this discussion.
The modification in the hydration kinetics of PHCM mixtures compared to that of
the control mixture can be ascribed to a number of effects. Adding pre-hydrated C3S
provides CH and CSH acts as nuclei for further CSH formation or for the conversion of
the ‘first-stage’ CSH (water impermeable product formed within the first few minutes of
hydration), into a ‘second-stage’ CSH (better permeable hydrate). Hence, this way
accelerates the hydration reactions, leading to abolishing the induction period. This was
found to be consistent with NMR results. Sets of NMR spectra for control and H2
samples at ages of 8, 10, 12 and 24 hours are shown in Fig. 6-12. The peaks
corresponding to CSH (Q1 and Q2) occur at about -79 and -84 ppm, respectively, similar
to previous findings (Fernandez et al., 2008, Al-Dulaijan et al., 1995, Cong and
Kirkpatrick, 1996). During the first hours of hydration, CSH formed with a
Chapter 6 206
predominantly protonated Q1 species, which represents the onset of SiO4 polymerization
(Johansson et al., 1999). As hydration progressed, increased amount of protonated Q2
species also became detectable in addition to Q1 (Lea, 1988). Comparing the NMR
spectra for control and H2, Q1 and Q2 had appeared, as early as 8 and 10 hrs for the H2
mixture and at 12 hrs for control mixture, respectively. This indicates an earlier onset of
polymerization and conversion of CSH from the first to the second stage for H2
compared to C, as shown in Fig. 6-12. These results are consistent with the early-age heat
evolution and compressive strength measurements.
(a) (a)
(a)
Figure 6-12: NMR spectra for a) C and b) H2 samples at different ages.
Chapter 6 207
(b)
(b)
Figure 6-12 Contd’: NMR spectra for a) C and b) H2 samples at different ages.
Moreover, added crystalline CH, besides acting as a passive internal restraint, it
enables further dissolution of C3S, renewed CSH formation (Tadros et al., 1976, Kondo
and Daimon, 1969). Furthermore, this reduction in the calcium and silicate ions in the
solution can result in the breakdown of the electrical double layer of calcium and silicate
ions formed on C3S during its initially dissolution, as indicated by the electrical double
layer theory. This can also lead to renewed acceleration of the hydration process (Lea,
1988, Tadros et al., 1976, Kondo and Daimon, 1969, Young et al., 1977). This is
depicted by the DSC results in Fig. 6-13, where enthalpies of CH (440°C) for the H2
were significantly higher than that of the C over the investigated period.
Chapter 6 208
0
10
20
30
40
2 4 6 8 12 16 18 24Age (hour)
Enth
alpy
(J/g
)
CH2
Figure 6-13: Enthalpy values for C and H2 mixtures during the first 24 hours.
Another mechanism emanates from the principle that silica fume (amorphous
silica) affects the rate of hydration during the early-age. Silica fume provides large
amounts of reactive siliceous surface, serving as a site for the early CSH precipitation.
Accordingly, the initial CSH layer will be formed mostly on silica surface rather than
developing a protective layer on the surface of the most reactive phases (i.e. C3S), thus
preventing a hydration delay (Korpa et al., 2008). Microstructure analysis confirms this
hypothesis as explained below. Furthermore, the pozzolanic reaction of silica fume with
CH (provided by PHCM) leads to an increase in the amount of developed CSH (less
protective than the initially formed CSH (McCurdy and Erno, 1970)) and the Q2 signal
which indicates a growth in the mean chain length of CSH phases (i.e. degree of
polymerization) (Schachinger et al., 2008). This is confirmed by the consumption of
Chapter 6 209
silica fume during the hydration period as detected by NMR results. Figure 6-12(b)
shows the gradual reduction in the silica fume NMR peak with time. Also, this early
pozzolanic reaction can be considered the reason for the high early-age compressive
strength (Al-Dulaijan et al., 1995).
The microstructure of H2 specimens was found to be uniformly developed in the
inter-granular space, as shown in Fig. 6-14(a). More anhydrous silica fume clusters and
large plate-like formations of CH can be observed clearly within the H2 matrix during
early-age (about 8 hours) compared to that at later ages (after 24 hours). This observation
emphasizes the role of PHCM in providing crystalline CH during early hydration as
mentioned before. In addition, the thin rim of CSH was not detected at that early-age.
Conversely, the H2 matrix at 24 hrs was found to be rich in dark CSH with a Ca/Si ratio
of about 0.86, as shown in Fig. 14(b). This reveals the effect of the pozzolanic reactions
of silica fume, which consumed the CH, producing less protective dark CSH (Scrivener,
2004) with low Ca/Si ratio range (0.83-1.5) (Lea, 1988), which allows more hydration
reactions to occur.
Chapter 6 210
(a)
(b)
Figure 6- 14: Microstructure of H2 mixtures after a) 8, and b) 24 hrs as in
backscattered electron microscopy, and energy dispersive X-ray element analysis.
Chapter 6 211
6.11. CONCLUSIONS
The main conclusions that can be drawn from this experimental investigation are the
following:
1) ent of the The addition of PHCM had a strong acceleration effect on the developm
internal microstructure leading to early setting and hardening process.
2) PHCM technique showed a high potential for reducing early-age shrinkage.
3) ad bearing structure Addition of PHCM increases contact points and acts as lo
providing an internal passive restraining system.
4) The higher the added PHCM, The higher was the restraining effect.
5) The PHCM technique initiates the concept of self-restraining concrete.
6) concrete in Concrete sustainability enhanced through using left-over and returned
produce self-accelerated concrete, thus preventing waste and disposal.
7) l for reducing autogenous Mixtures incorporating PHCM achieved a high potentia
shrinkage through providing internal passive restrains.
8) The PHCM showed an acceleration effect, which has a paramount potential,
particularly in the pre-cast industry. It resolves the two major drawbacks
associated with chloride accelerator; corrosion related to chlorides and increased
shrinkage. (A Comparison between PHCM and conventional chloride and non
chloride accelerators are shown in Appendix B).
Chapter 6 212
6.12. REFERENCES
ACI C
l-Dulaijan, S.U., Al-Tayyib, A-H.J., Al-Zahrani, M.M., Parry-Jones, G. and Al-Mana,
fume,” Journal of the American Ceramic Society, Vol. 78, No. 2, pp. 342-346.
Aitcin,
aroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafaï, N., (2006),
d Concrete Research, Vol. 31, No. 11, pp. 1587-1591.
3, pp. 423-429.
Bentz, Cement and Concrete Composites, Vol. 26, No. 6, pp. 677-685.
ommittee E-701, (2001). Cementitious Materials for Concrete, ACI Education Bulletin E3-01, American Concrete Institute, Farmington Hills, Michigan, 25 p.
AA.I., (1995), “29Si MAS-NMR study of hydrated cement paste and mortar made with and without silica
P.C., (1999), “Demystifying autogenous shrinkage,” Concrete International, Vol. 21, No. 11, pp. 54-56.
Barcelo, L., Boivin, S., Acker, P., Toupin, J. and Clavaud, B., (2001), “Early age
shrinkage of concrete: back to physical mechanisms,” Concrete Sciences and Engineering, Vol. 3, No. 10, 2001, pp. 85-91.
Barcelo, L., Moranville, M. and Clavaud, B., (2005), “Autogenous shrinkage of concrete:
a balance between autogenous swelling and self-desiccation,” Cement and Concrete Research, Vol. 35, No. 1, 2005, pp. 177-183.
B“Autogenous deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp. 123-136.
Bentur, A., Igarashi, S. and Kovler, K., (2001),“Prevention of autogenous shrinkage in
high strength concrete by internal curing using wet lightweight aggregates,” Cement an
Bentz, D.P., (2006), “Influence of shrinkage-reducing admixtures on early-age properties
of cement pastes,” Journal of Advanced Concrete Technology, Vol. 4, No.
D.P. and Jensen, O.M., (2004), “Mitigation strategies for autogenous shrinkage cracking,”
Bentz, D.P. and Peltz, M.A., (2008), “Reducing thermal and autogenous shrinkage
contributions to early-age cracking,” ACI Materials Journal, Vol. 105, No. 4, pp. 414-420.
Carde, C. and Francois, F., (1997), “Effect of the leaching of calcium hydroxide from
cement paste on mechanical and physical properties,” Cement and Concrete Research, Vol. 27, No. 4, pp. 539-550.
Chapter 6 213
Chikh, N., Cheikh-Zouaoui, M., Aggoun, S. and Duval, R., (2008), “Effects of calcium nitrate and triisopropanolamine on the setting and strength evolution of Portland cement pastes,” Materials and Structures, Vol. 41, No. 1, pp. 31-36.
y, Sweden.
o. 17, pp. 5772-5783.
nce and Technology, Trondheim, Norway.
Hansen of concrete,” Journal of American Ceramic Society, Vol. 70, No. 5, pp. 329-332.
HansenNo. 3, pp. 48-51.
.
rete”, Taylor, New York, pp. 143-154.
Holsch r, K., Dehn, F., Klotz, S. and Weiße, D., (2005), “Experimental
investigation on ultra high-strength concrete under concentrated loading,” In:
ncrete, Washington D.C., USA, Vol. 2, pp. 1145-1158.
Holt, E hesis, University of Washington, United States.
Cong, X. and Kirkpatrick, R.-J., (1996), “29Si MAS NMR study of the structure of
calcium silicate hydrate,” Advanced Cement Based Materials, Vol. 3, No. 3-4, pp. 144-156.
Esping, O., (2007), “Early-age properties of self-compacting concrete - Effects of fine
aggregate and limestone filler,” PhD Thesis, Department of Civil and Environmental Engineering, Chalmers University of Technolog
Fernandez, L., Alonso, C., Andrade, C. and Hidalgo, A., (2008), “The interaction of
magnesium in hydration of C3S and CSH formation using 29Si MAS-NMR,” Journal of Materials Science, Vol. 43, N
Fosså, K.T., (2001), “Slipforming of vertical concrete structures: Friction between
concrete and slipform panel,” PhD Thesis, the Norwegian University of Scie
, W., (1987), “Constitutive model for predicting ultimate drying shrinkage
, W., (2011), “Report on early-Age cracking: A summary of the latest document from ACI Committee 231,” Concrete International, Vol. 33,
Hammer, T.A., (1999a), “Early age chemical shrinkage and autogenous deformation of
cement paste,” In: Proceedings of the International Seminar on Self-desiccation and its importance in concrete technology, Lund, pp. 27-33
Hammer, T.A., (1999b), “Test methods for linear measurements of autogenous shrinkage
before setting,” Proceedings of the International Workshop “Autogenous Shrinkage of Conc
Henkensiefken, R., Briatka, P., Bentz, D.P., Nantung, T. and Weiss, J., (2010), “Plastic
shrinkage cracking in internally cured mixtures,” Concrete International, Vol. 32, No. 2, pp. 49-54.
emache
Proceedings of the Seventh International Symposium on Utilization of high-strength/ high performance co
.E., (2001), “Early-age autogenous shrinkage of concrete,” PhD T
Chapter 6 214
. and Leivo, M., (2004), “Cracking risks associated with early age shrinkage,” Cement and Concrete Composit
Holt, Ees, Vol. 26, No. 5, pp. 521-530.
ensen, O.M. and Hansen, P.F., (1995), “A dilatometer for measuring autogenous
ensen, O.M. and Hansen, P.F., (1996), “Autogenous deformation and change of relative
ensen, O.M. and Hansen, P. F., (1999), “Influence of temperature on autogenous
ensen, O.M. and Hansen, P.F., (2001), “Water-entrained cement-based materials: I.
Johanss and Ronin, V.,
(1999), “Kinetics of the hydration reactions in the cement paste with
stens, H., Van Gemart, A., Verboven, F. and Sellevold, E.J., (1996), “Total and
ondo, R. and Daimon, M., (1969), “Early hydration of tricalcium silicate: a solid
Korpa, A., Kowald, T. and Trettin, R., (2008), “Hydration behaviour, structure and
morphology of hydration phases in advanced cement-based systems containing
ronlöf, A., Leivo, M. and Sipari, P., (1995), “Experimental study on the basic
4.
Lea, F.M., (1988). Lea's Chemistry of Cement and Concrete, 4th Edition, (ed.) P.C. Hewlett, J. Wiley, New York, 1053 p.
Jdeformation in hardening Portland cement paste,” Materials and Structures, Vol. 28, No. 181, pp. 406-409.
Jhumidity in silica fume modified cement paste,” ACI Materials Journal, Vol. 93, No. 6, pp. 539-543.
Jdeformation and relative humidity change in hardening cement paste,” Cement and Concrete Research, Vol. 29, No. 4, pp. 567-575.
JPrinciples and theoretical background,” Cement and Concrete Research, Vol. 31, No. 4, pp. 647-654.
on, K., Larsson, C., Antzutkin, O.N., Forsling, W., Kota, H.R.
mechanochemically modified cement 29Si magic-angle-spinning NMR study,” Cement and Concrete Research, Vol. 29, No. 10, pp. 1575-1581.
Juexternal chemical shrinkage of low w/c ratio cement pastes,” Advances in Cement Research, Vol. 8, No. 31, pp. 121-126.
Kreaction with induction and acceleration periods,” Journal of the American Ceramic Society, Vol. 52, No. 9, pp. 503-508.
micro and nanoscale pozzolanic additives,” Cement and Concrete Research, Vol. 38, No. 7, pp. 955-962.
Kovler, K. and Zhutovsky, S., (2006), “Overview and future trends of shrinkage
research,” Materials and Structures, Vol. 39, No. 9, pp. 827-847.
Kphenomena of shrinkage and cracking of fresh mortar,” Cement and Concrete Research, Vol. 25, No. 8, pp. 1747-175
Chapter 6 215
., Guthrie, W.F. and Kacker, R., (1995), “A study of the reuse of plastic concrete
using extended set-retarding admixtures,” Journal of Research Lobo, C
of National Institute of Standards and Technology, Vol. 100, No. 5, pp. 575-589.
Ma, J.
al Symposium on Ultra High Performance Concrete, Germany, pp. 205-212.
McCur alcium silicate,” Canadian Journal of Chemistry, Vol. 48, No. 21, pp. 3291-3299.
Mehta, structure, Properties, and Materials. 3rd Edition, McGraw-Hill, New York, 659 p.
eville, A.M., (1996). Properties of Concrete. 4th Edition, Wiley, New York, 844 p.
Obla, K: Final report,” The RMC Research & Education Foundation, Project
05-13, 51 p.
Odler, hydration products,” Cement and
Concrete Research, Vol. 9, No. 2, pp. 239-248.
Powers Association Research & Development Laboratories, Vol. 4, No. 2, pp.
40-50.
Powers
rete Institute, October 1946–April 1947, Proceedings 43, Detroit, pp. 971-992.
Ramacnstruction Materials, Noyes & William Andrew
Publishing, Norwich, 655 p.
Roy, D rete Microstructure. SHRP-C340, Strategic Highway Research Program, 179 p.
Scrivenuantification,” Cement and Concrete
Composites, Vol. 26, No. 8, pp. 935-945.
, Orgass, M., Dehn, F., Schmidt, D and Tue, N.V., (2004), “Comparative investigations on ultra-high performance concrete with and without coarse aggregates,” Proceedings of the Internation
dy, K.G. and Erno, B.P., (1970), “The kinetics of hydration of tric
P.K. and Monteiro, P.J.M., (2006). Concrete: Micro
N
., Kim, H. and Lobo, C., (2007), “Crushed returned concrete as aggregates for new concrete
I. and Dörr, H., (1979), “Early hydration of tricalcium silicate I : Kinetics of the hydration process and the stoichiometry of the
, T.C., (1962), “A hypothesis on carbonation shrinkage,” Journal of the Portland Cement
, T.C. and Brownyard, T.L., (1947), “Studies of the physical properties of hardened portland cement paste,” Bulletin No. 22, 1948, Research Laboratories of the Portland Cement Association. Reprinted from Journal of the American Conc
handran, V.S., Haber, R.M., Beaudoin, J.J. and Delgado, A.H., (2002). Handbook of Thermal Analysis of Co
.M. and Idorn, G.M., (1993). Conc
er, K.L., (2004), “Backscattered electron imaging of cementitious microstructures: under- standing and q
Chapter 6 216
Schachinger, I., Hilbig, H. and Stengel, T., (2008), “Effect of curing temperature at an early age on the long-term strength development of UHPC,” In: Proceedings of the Second International Symposium on Ultra High-Performance Concrete, Kassel, Germany, pp. 205-212.
Schemmel, J.J., Ray, J.C. and Kuss, M.L., (1999), “Impact of shrinkage reducing
admixture on properties and performance of bridge deck concrete, high-performance concrete research to practice,” ACI, Special Publication, SP-189, pp. 367-386.
Struble, L., Skalny, J. and Mindess, S., (1980), “A review of the cement-aggregate bond,”
Cement and Concrete Research, Vol. 10, No. 2, pp. 277-286. Tadros, M.E., Skalny, J. and Kalyoncu, R.S., (1976), “Early hydration of tricalcium
silicate,” Journal of the American Ceramic Society, Vol. 59, No. 7-8, pp. 344-347.
Tazawa, E., (1999). Autogenous Shrinkage of Concrete. In: Proceedings of the
International Workshop organized by the Japan Concrete Institute, Taylor, New York.
Tazawaa, E. and Miyazawaa, S., (1995a), “Influence of cement and admixture on
autogenous shrinkage of cement paste,” Cement and Concrete Research, Vol. 25, No. 2, pp. 281-287.
Tazawa, E. and Miyazawa, S., (1995b), “Experimental study on mechanism of
autogenous shrinkage of concrete,” Cement and Concrete Research, Vol. 25, No. 8, pp. 1633-1638.
Tazawa, E. and Miyazawa, S., (1999), “Effect of constituents and curing condition on
autogenous shrinkage of concrete,” In: Autogenous Shrinkage of Concrete: Proceedings of the International Workshop, Eiichi Tazawa (ed.), Taylor & Francis, pp. 269-280.
Wittmann, F.H., (1976), “On the action of capillary pressure in fresh concrete,” Cement
and Concrete Research, Vol. 6, No. 1, pp. 49-56. Xiao, L. and Li, Z., (2008), “Early-age hydration of fresh concrete monitored by non-
contact electrical resistivity measurement,” Cement and Concrete Research, Vol. 38, No. 3, pp. 312-319.
Young, J.F., Tong, H.S. and Berger, R.L., (1977), “Compositions of solutions in contact
with hydrating tricalcium silicate pastes,” Journal of the American Ceramic Society, V. 60, No. 5-6, pp. 193-198.
Chapter 7 217
CHAPTER SEVEN
EVALUATION OF SELF-RESTRAINING SHRINKAGE TECHNIQUE COMPARED TO CONVENTIONAL
METHODS IN UHPC: INDIVIDUALLY AND COMBINED*
In the previous chapter (Chapter 6), the concept of developing self-restraining concrete
was introduced. Using partially hydrated cementitious materials (PHCM) had shown a
high potential to improve the early-age properties of UHPC. In addition, it provides an
environmental and economic technique for producing precast and cast in place self-
restraining shrinkage concrete, through inducing internal passive restraint within fresh
concrete. In this chapter, the efficiency of PHCM as a shrinkage mitigation method was
evaluated in comparison to conventional techniques including, shrinkage-reducing
admixture (SRA) and/or a superabsorbent polymer (SAP). In addition, the combined
effects of PHCM, SRA and SAP techniques in reducing early-age shrinkage were
investigated.
7.1. INTRODUCTION
General literature about UHPC characterizations, potential applications, shrinkage
problems and different mitigation strategies has been given in proceeding chapters 3 and
4. It has been outlined that the ultimate mechanical properties and enhanced durability
make it a promising material for many construction applications. However, its high self-
desiccation shrinkage and potential of early-age cracking may defeat its purposes. Hence,
*A version of this chapter has been submitted for review to Cement and Concrete Research.
Chapter 7 218
several shrinkage mitigation methods were proposed. Among these methods, SRA is the
most commonly used method, while SAP represents an enhanced version of internal
curing technique.
In addition to the shrinkage mitigation strategies described above, various
concrete components can provide passive internal restraint that resists autogenous
shrinkage and reduces the measured physical deformations (Bentz and Jensen, 2004).
Chief amongst these are aggregate particles (Hobbs, 1974, Hansen and Nielsen, 1965). At
the micro-level, a passive restraint can be provided by several components including un-
hydrated cement particles, hydrated cement paste skeleton (Aitcin, 1999) and other
hydration products (e.g. calcium hydroxide (CH) micro-crystals) (Jensen and Hansen,
1996).
Therefore, it was proven in Chapter 6 that adding PHCM can provide an internal
restraint system which varies in size from as small as CH micro-crystals up to networks
of hydration products and clusters. These clusters can provide load-bearing mechanisms
at the micro-structural level to resist contracting forces in fresh concrete. Besides, using
PHCM has two major positive environmental and economical benefits through
eliminating, or greatly reducing wastage of concrete, which consequently can lead to cost
reduction. Hence, the focus of this chapter is to investigate the effectiveness of PHCM as
a passive internal restraint system to reduce early-age shrinkage strains in comparison to
SRA and SAP.
Chapter 7 219
7.2. RESEARCH SIGNIFICANCE
Early-age shrinkage cracking is a considerable problem for concrete structures. Several
shrinkage mitigation techniques have been proposed to reduce shrinkage and to avoid
crack formation. However, such mitigation methods can adversely affect other key
properties of concrete inducing difficulties in controlling consistency, retardation of
hydration reactions, reduction in mechanical properties, etc. Adding partially hydrated
cementitious materials (PHCM) to the fresh concrete during mixing had shown a high
potential to reduce early-age shrinkage. This chapter confirms the efficacy of PHCM as a
new environmental-friendly shrinkage mitigation technique compared to SRA and SAP.
PHCM mitigates undesirable behaviour induced by SRA and/or SAP, including setting
time delays and reduction in mechanical properties. Hence, unused/returned concrete can
be recycled in new mixtures, individually or combined with other shrinkage mitigation
techniques, to enhance mechanical strength and resistance to shrinkage, and to mitigate
set retardation effects.
7.3. EXPERIMENTAL PROGRAM
The experimental program aims at gaining an understanding of the mechanisms involved
in using PHCM. This is of primary importance in order to enable the production of
shrinkage self-restraining concrete. In this study, monitoring the hydration process
(degree of hydration, heat of hydration, Vicat setting time) and characterization of the
shrinkage behaviour have been carried out on concrete mixtures incorporating PHCM,
SRA and SAP. The chapter consists of two parts. Part I was devoted to investigating the
effects of PHCM addition on the early-age properties of control mixtures with and
without SRA and/or SAP. The second part focused on validating the addition of PHCM
Chapter 7 220
as a shrinkage mitigation technique and evaluating its efficiency compared to that of SRA
and SAP. Combined effects of the tested shrinkage mitigation techniques including
PHCM, SRA, and SAP were also investigated. Mixtures containing a moderate dosage of
PHCM (i.e. 33%) were selected for comparison in this second part. All tests were
conducted on UHPC specimens without heat curing in order to explore realistic effects
that govern UHPC shrinkage in structural elements cast in situ.
7.3.1. Materials and Mixture Proportions
The materials used in this chapter were similar to that used in Chapter 4 (refer to Section
4.4.1). The chemical and physical properties of the used binders have been given in
Chapter 3 (Table 3-1). The selected composition of the control and PHCM mixture are
shown in Chapter 6 (refer to Table 6-1). The characteristics of the tested mixtures are
shown in Table 7-1.
Table 7-1: Tested mixtures.
Mixture PHCM(%)* SRA (%)** SAP (%)**
C ‐‐‐‐ ‐‐‐‐ ‐‐‐‐
H1 25.0 ‐‐‐‐ ‐‐‐‐
H2 33.0 ‐‐‐‐ ‐‐‐‐
H3 50.0 ‐‐‐‐ ‐‐‐‐
CR2 ‐‐‐‐ 2.0 ‐‐‐‐
CS ‐‐‐‐ ‐‐‐‐ 0.6
CR2S ‐‐‐‐ 2.0 0.6
H2R2 33.0 2.0 ‐‐‐‐
H2S 33.0 ‐‐‐‐ 0.6
H2R2S 33.0 2.0 0.6 *Added as % of the batch weight ** Added as % by mass of cement
Chapter 7 221
7.3.2. Preparation of Test Specimens and Testing Procedures
The experimental methods used in this chapter are the cubic compressive strength test,
Semi-adiabatic calorimetry, TGA, measurements of total free shrinkage, setting time and
the corrugated tube protocol for measuring autogenous shrinkage. All experimental
methods were previously explained in chapters 3 and 6.
7.4. RESULTS AND DISCUSSION
7.4.1. Setting Time and Early-Age Compressive Strength
Regardless of the added amount of PHCM, the PHCM technique significantly affected
the development of early-age compressive strength. All UHPC mixtures incorporating
PHCM consistently produced higher early-age compressive strength compared to that of
the control mixture as discussed in chapter 6 (see Fig. 6-5(b)).
On the other hand, the mixture incorporating 2% SRA (CR2) achieved a lower
early-age compressive strength compared to that of the C (Fig. 7-1(a)). This can be
attributed to the retardation effect induced by SRA, in agreement with previous work (He
et al., 2006). The mixture incorporating PHCM and SRA (H2R2) achieved higher
compressive strength than that of the CR2 mixture, indicating that the accelerated
hydration induced by PHCM addition offset the early-age compressive strength reduction
induced by SRA. For instance, H2R2 exhibited a compressive strength of about 1.1 MPa
at the age of 6 hours, while CR2 did not exhibit any strength at that age (Fig. 7-1(a,b)).
The early-age compressive strength of the mixture incorporating SAP (CS) was
lower than that of the C (Fig. 7-1(a)). This can be ascribed to the higher water content of
Chapter 7 222
the CS mixture, leading to higher porosity and consequently lower strength (Jensen and
Hansen, 2002). Adding PHCM to the CS mixture improved its early-age compressive
strength. This can be explained as follows: the higher water content imparted by SAP
provided enhanced conditions for hydration reactions to progress. SAP particles release
their entrained water and occupy a lower volume (Wang et al., 2009), leading to more
space for hydration products to form. At such a very low w/c, the availability of space for
hydration products to form is a main restriction on hydration development (Acker, 2004).
Hence, enhanced curing conditions and availability of space along with accelerated
hydration reactions induced by PHCM addition produced a well developed
microstructure and consequently higher early-age strength was achieved.
The mixture incorporating both SRA and SAP (CR2S) exhibited a lower early-
age compressive strength than that of the CS; reflecting the set retarding effect induced
by SRA. However, the CR2S mixture exhibited higher early-age strength compared to
that of the CR2. This can be ascribed to the lower SRA concentration in CR2S as a result
of the higher water content (Rajabipour et al., 2008) and the ability of SAP particles to
absorb SRA (Bentz, 2005). Adding PHCM to the CR2S mixture enhanced its early-age
compressive strength by about 26% at 24 hours (Fig. 7-1(b)). This improvement exhibits
the resultant of conflicting effects of SRA, SAP and PHCM on the early-age strength
developments.
Chapter 7 223
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Com
pres
sive s
treng
th (M
Pa)
CCR2CSCR2S
(a)
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Com
pres
sive s
treng
th (M
Pa)
CR2SH2R2H2SH2R2S
(b)
Figure 7-1 : Early-age compressive strength development for mixtures
incorporating a) SRA and/or SAP and b) combined shrinkage mitigation
techniques.
Chapter 7 224
Adding SRA tends to prolong the setting time compared to that of the control
mixture, while adding SAP tends to shorten it slightly owing to the better curing
condition offered by the SAP entrained water and consequently the better progress of
hydration reactions. Adding PHCM shortened the setting time significantly compared to
that of the C mixture as shown in Table 7-2. The higher the PHCM addition, the shorter
was the measured setting time. These results confirm the acceleration of hydration
reactions imparted by the PHCM technique. This acceleration effect offsets the
undesirable behaviour induced by SRA and/or SAP. For instance, it mitigates the
retardation effect induced by SRA addition, leading to enhanced early-age compressive
strength (Fig. 7-1(b)).
Table 7-2 : Setting time results for the tested mixtures.
Setting Time Initial Final Mixture
Hours Minutes Hours Minutes C 5 25 6 10 H1 4 45 5 10 H2 3 05 4 00 H3 2 35 3 20 CR2 6 17 7 10 CS 5 08 5 50 CR2S 5 55 6 35 H2R2 5 35 6 15 H2S 4 45 5 30 H2R2S 5 00 5 53
7.4.2. Degree of Hydration
The early-age compressive strength development of concrete mixtures greatly depends on
the degree of hydration achieved (Xiao and Li, 2008). Shrinkage mitigation techniques
Chapter 7 225
usually affect the rate of hydration reactions. Accordingly, the degree of hydration and
amount of BW will be changed (Bentz et al., 2001a). TGA was used to investigate the
hydration evolution. The measured degree of hydration (represented by the relative BW)
versus age is shown in Fig. 7-2(a,b).
Adding PHCM accelerated the rate of hydration reactions, leading to a higher
degree of hydration. The higher the added dosage of PHCM, the higher was the degree of
hydration achieved as shown in Fig. 7-2(a). This is in agreement with previous early-age
compressive strength results (see Fig. 6-5(b)). Adding SRA retarded the hydration
reactions leading to a lower degree of hydration compared to that of the C mixture (Fig.
7-2(a)). Conversely, the CS mixture had relatively higher water content, thus achieving a
higher degree of hydration compared to that of the C mixture (Jensen and Hansen, 2001,
Jensen and Hansen, 2002). Despite its higher degree of hydration, the CS mixture
exhibited a lower early-age compressive strength compared to that of the C mixture (Fig.
7-1(a)) as a result of its higher porosity. Analysis of the incremental pore size distribution
data from MIP tests at age 24 hrs showed that SAP specimens had a higher proportion of
voids compared to that of the control specimens. The percentage of voids, classified
according to the International Union of Pure and Applied Chemistry system (IUPAC)
(IUPAC, 1972) was about 4.7% and 2.2% for mixtures CS and C, respectively. This can
be explained as follows: in the mixture without SAP, a lower amount of hydration is
required to gain strength since the space between solid particles and the porosity to be
filled by hydration products are relatively lower than that of the SAP mixture.
Conversely, in the SAP mixture, a higher degree of hydration is achieved while the
volume of pores to be filled increases, resulting in a higher net porosity. This is consistent
Chapter 7 226
with previous findings which indicate that porosity has a more dominate effect on the
achieved strength than that of the degree of hydration (Odler and Rößler, 1985, Mikhail
et al., 1977). The mixture incorporating both SRA and SAP (CR2S) did not exhibit a
significant improvement in degree of hydration; however, its degree of hydration was
higher than that of the CR2 mixture (Fig. 7-2(a,b)). Adding PHCM to the CR2 mixture
improved its degree of hydration significantly. After 24 hours, mixture H2R2 had about
18% higher BW than that of the CR2 mixture. This is ascribed to the hydration
acceleration effect induced by PHCM. On the other hand, adding PHCM to the CS
mixture resulted in the highest degree of hydration compared to that of the other
mixtures, indicating a considerable progress of hydration reactions (Fig. 7-2(b)). This can
be ascribed to the coupled effects of adequate curing conditions provided by SAP
particles and the acceleration effect of PHCM addition. Similar to the trend of C and CS
mixtures, the H2S mixture showed lower compressive strength than that of H2 mixture.
Adding SRA to the mixture incorporating PHCM and SAP (H2S) led to a reduction in
the amount of the BW (i.e. degree of hydration) compared to that of the H2S mixture,
which is expected due to the retardation effect imparted by SRA.
Chapter 7 227
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0 5 10 15 20 25Age (hours)
Rel
ativ
e B
W (g
/g b
inde
r)
CCR2CSH1H2H3
(a)
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0 5 10 15 20 25Age (hours)
Rel
ativ
e B
W (g
/g b
inde
r)
CR2SH2R2SH2SH2R2
(b)
Figure 7-2: Degree of hydration development for mixtures incorporating a) single
and b) combined shrinkage mitigation techniques.
Chapter 7 228
7.4.3. Heat of Hydration
The heat of hydration for mixtures incorporating PHCM has been discussed in chapter 6
(see Fig. 6-8). Adding PHCM effectively diminished the induction period, leading to a
continuous progress of hydration reactions and consequently shorter setting time and
more advanced development of microstructure.
It can be noted that adding SRA retarded hydration reactions and reduced the
liberated heat compared to that of mixture C, as shown in Fig. 7-3(a). This is consistent
with previous results (Eberhardt and Kaufmann, 2006). Adding PHCM to the CR2
mixture resulted in a higher heat of hydration compared to that of the original CR2
mixture. Figure 7-3(b) illustrates the synergetic effect of PHCM and SRA, which
represents the net result of two conflicting effects: acceleration and retardation. Adding
PHCM accelerated the hydration process and shifted the hydration curve upward and to
the left (compared to that of the CR2 mixture). Concurrently, SRA limited the
acceleration effect induced by PHCM addition, resulting in a slightly lower hydration
peak at a later time of about 1 hour with respect to that of the H2 mixture.
The mixture incorporating SAP (CS) exhibited a different trend than that of
mixture C. It initially exhibited a lower temperature peak. This can be ascribed to the
high volume fraction of water in the CS mixture, which led to higher heat capacity
compared to that of the C mixture (Bentz, 2008). After about 24 hours, another peak
occurred, which indicates further hydration. This can be ascribed to the further hydration
reactions between the un-hydrated binder and water released from SAP particles. The
mixture incorporating PHCM and SAP (H2S) exhibited a higher temperature peak
compared to that of the H2 mixture. Moreover, there was no indication of a second peak
Chapter 7 229
as in the CS mixture. This can be ascribed to the early accelerated hydration induced by
PHCM, which consumed the majority of the available water (i.e. original mixing water
and water entrained by SAP particles) leading to a single temperature peak.
The combined effect of SRA and SAP (mixture CR2S) slightly affected the
hydration curve with respect to that of the CR2 and CS mixtures. The set retardation
effect of SRA shifted the temperature peak downward and to a later time by
approximately 1 hour with respect to that of the CS mixture. Moreover, similar to the
case of the CS mixture, a small secondary peak occurred. The mixture incorporating
PHCM, SRA and SAP (H2R2S) showed a combined trend to that of the CS, CR2S and
H2 mixtures. The temperature maximum shifted upward and to an earlier time by
approximately 3 hours compared to that of the CR2S mixture as a result of the hydration
acceleration effect induced by PHCM. However, H2R2S exhibited a lower temperature
peak compared to that of the H2 mixture, which can be ascribed to the SRA hydration
retardation effect and higher heat capacity (as a result of higher water content) (Fig. 7-
3(c)). Moreover, a secondary temperature peak occurred at about 8 hours earlier than that
of the CS mixture (Fig. 7-3(a,b)). This indicates that SAP released its entrained water
earlier, reflecting the higher water demand induced by the accelerated hydration.
Chapter 7 230
20
25
30
35
40
45
0 6 12 18 24 30 36 42 48Age (hours)
Tem
pera
ture
(°C
)
CCR2CSCR2S
(a)
20
25
30
35
40
45
0 6 12 18 24 30 36 42 48Age (hours)
Tem
pera
ture
(°C
)
H2H2R2H2SH2R2S
(b)
Figure 7-3: Heat of hydration development for mixtures incorporating a) SRA
and/or SAP and b) combined shrinkage mitigation techniques.
Chapter 7 231
7.4.4. Shrinkage and Mass Loss
The mass loss and shrinkage curves of mixtures incorporating SRA, SAP and/or PHCM
are shown in Fig. 7-4 and 7-5. The measured total shrinkage includes both drying and
autogenous shrinkage (thermal deformation can be ignored due to the small cross-section
of the samples (Baroghel-Bouny et al., 2006)). For very low w/c (Tazawa and Miyazawa,
1999), the contribution of autogenous shrinkage to the total shrinkage can be comparable
to that of drying shrinkage.
0
0.5
1
1.5
2
2.5
24 48 72 96 120 144 168Age (hours)
Mas
s lo
ss (%
)
CCR2CSH1H2H3
(a)
Figure 7-4: Mass loss for mixtures incorporating a) single and b) combined
shrinkage mitigation techniques.
Chapter 7 232
0
0.5
1
1.5
2
2.5
24 48 72 96 120 144 168Age (hours)
Mas
s lo
ss (%
)
CR2SH2R2H2SH2R2S
(b)
Figure 7-4 Contd’: Mass loss for mixtures incorporating a) single and b) combined
shrinkage mitigation techniques.
-600
-500
-400
-300
-200
-100
024 48 72 96 120 144 168
Age (hours)
Shrin
kage
(µε)
CCR2CSH1H2H3
(a)
Figure 7-5: Shrinkage development for mixtures incorporating a) single and b)
combined shrinkage mitigation techniques.
Chapter 7 233
-600
-500
-400
-300
-200
-100
024 48 72 96 120 144 168
Age (hours)
Shrin
kage
(µε)
CR2SH2SH2R2H2R2S
(b)
Figure 7-5 Contd’: Shrinkage development for mixtures incorporating a) single and
b) combined shrinkage mitigation techniques.
Incorporating SRA slightly reduced the mass loss with respect to that of the C
mixture, yet it induced a significant reduction in the total shrinkage, in agreement with
previous findings (Ichinomiya et al., 2005, Bentz, 2006, He et al., 2006). Figure 7-6
demonstrates the relationship between a given mass loss and the corresponding
shrinkage. The lower slope for the initial linear part and the second part of the mass loss-
shrinkage curve for mixture CR2 compared to that of mixture C is consistent with the
mechanisms of reduction in surface tension induced by SRA (Kovler and Bentur, 2009).
As a result, lower capillary stresses developed, resulting in lower total shrinkage for a
given mass loss.
Chapter 7 234
-600
-500
-400
-300
-200
-100
00.0 0.5 1.0 1.5 2.0 2.5
Mass loss (%)
Shrin
kage
(µε)
C ; (Si =170, Ss=561)CR2; (Si =77, Ss=391)CS ; (Si =135, Ss=417)H1 ; (Si =173, Ss=470)H2 ; (Si =168, Ss=454)H3 ; (Si =161, Ss=431)
(a)
Slope Initial Second
C -170 -560 CR2 -77 -391 CS -135 -417 H1 -173 -470 H2 -168 -454 H3 -161 -431
-600
-500
-400
-300
-200
-100
00.0 0.5 1.0 1.5 2.0 2
Mass loss (%)
Shrin
kage
(µε)
.5
CR2S; (Si=116, Ss=388)H2R2, (Si=137, Ss=328)H2S; (Si=165, Ss=355)H2R2S; (Si=142, Ss=273)
(b)
Slope Initial Second
CR2S -116 -338 H2R2 -165 -355 H2S -137 -328
H2R2S -142 -273
Figure 7-6: Mass loss- shrinkage relationship for mixtures incorporating a) single
and b) combined shrinkage mitigation techniques.
Chapter 7 235
This reduction in capillary stresses also explains the lower autogenous shrinkage
of mixture CR2 compared to that of mixture C (about 59% reduction at 7 days) as shown
in Fig. 7-7(a). Moreover, SRA is believed to reduce the drop in internal relative humidity
of specimens, leading to lower self-desiccation and autogenous shrinkage (Bentz et al.,
2001b). A few hours after final setting, CR2 specimens showed an expansion of
approximately 85 µε compared to only about 21 µε for C specimens (Fig. 7-7(a)). This
expansion along with the reduction in autogenous shrinkage induced by SRA resulted in
lower net shrinkage compared to that of C specimens (Weiss et al., 2008).
Conversely, CS specimens exhibited higher mass loss (about 35% at 7 days)
compared to that of C specimens, while the measured shrinkage at the same age was
increased by only 12%. Consequently, the mass loss-shrinkage curve for CS mixture
(Fig. 7-6(a)) exhibited lower slope compared to that of the C mixture. This can be
explained by the specific shrinkage mitigation mechanism of SAP. The higher water
content (due to added entrained water) initially results in higher mass loss (Mönnig and
Lura, 2007). Concurrently, SAP particles release water and occupy a smaller volume,
leading to additional coarse pores from which water first evaporates, regardless of its
location from the drying surface (Bentz et al., 2001a). Such water evaporation results in
lower capillary stresses compared to that in the C mixture having a finer porosity. In
addition, releasing the entrained water from SAP particles likely resulted in expansion
(Jensen and Hansen, 2002). Different phenomena are believed to induce this expansion,
including swelling of cement gel due to water absorption (Jensen and Hansen, 2002), and
capillary distension as a result of capillary vapour pressure increase leading to capillary
surface tension relaxation and consequently lower capillary stresses (Kovler, 1996).
Chapter 7 236
Furthermore, autogenous shrinkage and expansion occur simultaneously during early-
age, resulting in lower net autogenous shrinkage. This is consistent with the lower
autogenous shrinkage exhibited by CS specimens compared to that of C specimens (22%
reduction) (Fig. 7-7(a)) and is in agreement with previous findings (Baroghel-Bouny et
al., 2006, Kamen et al., 2008)
PHCM mixtures showed a slight reduction in the mass loss and measured total
shrinkage compared to that of the C mixture (Figs. 7-4 and 7-5). For instance, the H3
mixture had the lowest mass loss among the other PHCM mixtures, which was about
10% lower than that of the C mixture. Increasing the amount of added PHCM reduced the
measured shrinkage. This can be ascribed to the passive internal restraint system (i.e.
hydration micro-crystals (Aitcin, 1999, Jensen and Hansen, 1996)), provided by PHCM
addition. The hypothesis that CH micro-crystals provide restraint was proposed by
(Jensen and Hansen, 1996) based on results reported in (Carde and Francois, 1997,
Powers, 1962) which indicated a significant effect of leaching of CH micro-crystals on
deformation properties. Adding PHCM accelerates the hydration process leading to two
conflicting mechanisms: higher rate of autogenous shrinkage and faster structuring of a
stiff hydrated skeleton. This internal skeleton can withstand compressive stresses induced
by capillary stress due to either drying or self-desiccation (Aitcin, 1999, Acker, 2004,
Anna et al., 1995). Consequently, PHCM mixtures had a lower shrinkage, leading to a
lower mass loss-shrinkage ratio as shown in Fig. 7-6(a). PHCM mixtures exhibited slope
of the initial part of the mass loss-shrinkage curves comparable to that of the C mixture,
reflecting the similarity of the large pore structure (Kovler and Bentur, 2009), while the
Chapter 7 237
slopes of the second part were lower, indicating that the same water loss resulted in lower
shrinkage strain.
The relatively stiffer skeleton and passive internal restraint system induced by
PHCM can explain the significant reduction in autogenous shrinkage of PHCM mixtures
compared to that of the C mixture (Fig. 7-7(a)). The reduction in the measured
autogenous shrinkage after 24 hours versus that of C was about 10% for H1, 37% for H2,
and 57% for H3.
-800
-700
-600
-500
-400
-300
-200
-100
00 24 48 72 96 120 144 168
Age (hours)
Aut
ogen
ous
shrin
kage
(µε)
CCR2CSH1H2H3
(a)
Figure 7-7: Autogenous shrinkage for mixtures incorporating a) single and b)
combined shrinkage mitigation techniques.
Chapter 7 238
-800
-700
-600
-500
-400
-300
-200
-100
00 24 48 72 96 120 144 168
Age (hours)
Aut
ogen
ous
shrin
kage
(µε)
CR2SH2R2H2SH2R2S
(b)
Figure 7-7 Contd’: Autogenous shrinkage for mixtures incorporating a) single and
b) combined shrinkage mitigation techniques.
Figure 7-7(b) shows the measured shrinkage for mixtures incorporating more
than one shrinkage mitigation technique. Generally, the shrinkage behaviour of these
mixtures depended on the interaction between the different shrinkage mitigation
mechanisms of the combined techniques. This probably can lead to affirmative and/or
negative effects on the final shrinkage behaviour.
As mentioned earlier, adding SRA leads to a lower shrinkage as a result of
reducing capillary stresses, inducing early expansion, and retarding hydration reactions,
which effectively reduces the rate of self-desiccation (i.e. autogenous shrinkage)
development. Therefore, adding SRA to SAP and PHCM mixtures reduced the total and
autogenous shrinkage significantly with about 19% and 25% with respect to SAP mixture
Chapter 7 239
values, and by 26% and 17% with respect to PHCM values, respectively. In fact, the
early expansion had a major contribution to the achieved reduction in the net autogenous
shrinkage. Mixture incorporating SAR and SAP exhibited higher early expansion (about
74 µm) which is about one order of magnitude of that of SAP mixtures without SRA (32
µm). Moreover, adding SRA to PHCM led to an early-age expansion, distinct from
mixture incorporating PHCM only, which showed a plateau in shrinkage development
during the same period. This indicates that the early expansion effect of SRA was a
dominate factor in the resultant shrinkage behaviour.
Conversely, adding either SAP or PHCM to CR2 mixture increases the measured
shrinkage compared to mixture incorporating SRA only. This can be attributed to the
reduction in the SRA concentration due to the higher water content (Acker, 2004) and the
ability of SAP particles to absorb SRA (Bentz, 2005). Moreover, SRA hinder CH
formation (Matlese et al., 2005) which results in a lower amount of micro restraining
crystals compared to PHCM mixture as shown in TGA curves (Fig. 7-8). PHCM mixture
incorporating SRA showed an additional peak at about 340°C and smaller amount of CH
decomposition compared to that of PHCM mixture without SRA. Previous studies (Sant,
2009) reported a reduction in the SRA peak with time, which was attributed to the
reduction in SRA concentration in the pore solution due to being uptake by hydration
products. This can explain the difference in the 12 and 24 hrs TGA results, where the
peak corresponding to SRA was disappeared for H2R2 mixture at 24 hrs and another
peak seems to detected around 800°C, which is correspond to SRA taken in the CSH gel
(Beaudoin et al., 2008).
Chapter 7 240
-0.34
-0.28
-0.22
-0.16
-0.10
-0.04
0.020 200 400 600 800 1000 1200
Temperature (°C)
Diff
eren
tial W
eigh
t Los
s (%
/°C)
H2-12hrsH2-24hrsH2R2-12hrsH2R2-24hrs
SRA decomposition (~320 °C)
SRA in CSH gel decomposition (~800°C)CH decomposition
(~400-450°C)
Figure 7-8: TGA curves at different ages for PHCM mixture with and without SRA.
On the other hand, adding PHCM to SAP mixture showed about 16% and 11%
lower mass loss and total shrinkage, respectively compared to that the SAP mixture, and
slightly higher shrinkage compared to that of the PHCM mixture. This can be explained
as follows: the addition of PHCM accelerates hydration reactions, which increases water
consumption and the rate of self-desiccation. Hence, the water stored inside the SAP
particles will be released earlier to compensate for the water deficit. This leads to less
evaporable water, more hydration products and stronger restraining skeleton in the H2S
mixture compared to that of the CS mixture. This can be seen in TGA curves, where the
peak of evaporable water (60-200°C) was higher in SAP compared to that of H2S
mixture, while the CH peak (thus amount) was lower (Fig. 7-9(a)). This was also
confirmed by DSC results, where CS mixtures showed lower CH enthalpy compared to
that of mixture incorporating both PHCM and SAP (Fig. 7-9(b)). Hence, the absence of
Chapter 7 241
the early expansion in P-SAP specimens compared to that of the SAP specimens (Fig. 7-
7(a,b)) can be ascribed to the higher hydration progress and consequently higher
autogenous shrinkage that likely offset the expansion induced by the released water. This
is in agreement the previous heat of hydration results of H2S mixture (higher temperature
peak and the absent of the second peak (Fig. 7-3(b))).
-0.34
-0.28
-0.22
-0.16
-0.10
-0.04
0.020 200 400 600 800 1000 1200
Temperature (°C)
Diff
eren
tial W
eigh
t Los
s (%
/°C)
CS-12hrsCS-24hrsH2S-12hrsH2S-24hrs
CH decomposition (~400-450°C)
(a)
Figure 7-9: Thermal analysis at different ages for PHCM mixture with and without
SAP a) TGA and b) CH content.
Chapter 7 242
18.03
45.3242.12
32.81
0
10
20
30
40
50
SAP PHCM-SAP SAP PHCM-SAP
Enth
alpy
(J/g
)
12 hrs
24 hrs
(b)
Figure 7-9 Contd’: Thermal analysis at different ages for PHCM mixture with and
without SAP a) TGA and b) CH content.
As expected, combining the three mitigation techniques showed a trend that can
be a considered as the resultant of the different involved mechanisms. Adding SRA
retarded hydration and induced higher early expansion (about 70 µm) compared to
mixtures incorporating both PHCM and SAP, which showed a plateau in stead of
expansion during the same period (Fig. 7-7(b)). Adding PHCM provided additional CH
to the mixture, which enhanced the passive retraining system and eliminate the hindering
of CH formation induced by SRA. This was confirmed based on DSC results, where the
CH enthalpy in CR2S mixture, detected as early as 3 hrs from adding the mixing water,
was increased with the addition of PHCM to about the double (i.e. 3.104 J/g).
Chapter 7 243
The mass loss-shrinkage curves for mixtures incorporating combined shrinkage
mitigation techniques are shown in Fig. 7-6(b). Adding SRA to SAP and/or PHCM
mixtures reduced the initial slope. This is expected due to the additional shrinkage
reduction induced by SRA. Adding SAP to the H2 mixture did not cause a significant
change in the initial slope. This is consistent with the earlier explanation of the synergetic
effect of PHCM and SAP.
Moreover, mass loss-shrinkage curves second slopes for all mixtures
incorporating more than one mitigation technique were lower than those of mixtures
incorporating one mitigation technique (Fig. 7-6(b)), indicating higher efficiency than
their individual effects. For instance, the three mitigation techniques individually reduced
the second slope with an average of 25% compared to 51% for the H2R2S mixture with
respect to mixture C value.
7.5. STATISTICAL ANALYSIS FOR EFFECT OF PHCM ADDITION ON
SHRINKAGE
Analysis of variance (ANOVA) was used to analyze the experimental data. To
investigate whether an experimental variable (e.g. PHCM addition) is statistically
significant, an F value is determined as the ratio of the mean squared error between
treatments (e.g. different PHCM addition ratio) to that of within treatments (due to using
replicates rather than testing only one specimen). This value is then compared to a
standard (critical) F value of an F-distribution density function obtained from statistical
tables based on the significance level (α) and the degrees of freedom of error determined
from the number of treatments and observations in an experiment. Exceeding the critical
Chapter 7 244
value of an F-distribution density function reflects that the tested variable affects the
mean of the results (Montgomery, 2009).
ANOVA at a significance level α = 0.05, showed that variation in the addition
rate of PHCM had an insignificant effect on the mean of total shrinkage measured after
the first 24 hours. The calculated F value of 0.892 for the total shrinkage results was
lower than the corresponding critical F value of 4.07 (F0.05,3,8). Conversely, variation in
the addition rate of PHCM showed a significant effect on the mean autogenous shrinkage
measured from the final setting time; the associated value was 60.26, which is
significantly larger than the corresponding critical F value (F0.05,3,8).
According to previous studies (Aitcin, 1999, Zhang et al., 2003), measuring the
total shrinkage after 24 hours leads to overlooking the autogenous deformation
contribution. Hence, the total measured shrinkage can be re-evaluated taking into account
the autogenous shrinkage measured using the corrugated tube from the time of final
setting until the beginning of drying at 24 hours (Weiss et al., 2008). ANOVA for the
total measured shrinkage after adding autogenous shrinkage showed that the variation in
the addition rate of PHCM had a significant effect on the mean of the total shrinkage
results. The calculated F value was 73.03, which is significantly larger than the
corresponding critical F value (F0.05,3,8). This emphasizes the significant role of
autogenous shrinkage. Indeed, delays in measuring shrinkage can result in
underestimating the actual deformation behaviour, in agreement with previous findings
(Aitcin, 1999, Zhang et al., 2003).
Chapter 7 245
7.6. CONCLUSIONS
The main conclusions that can be drawn from this experimental investigation are:
1) UHPC mixtures incorporating PHCM achieved higher early-age compressive
strength results than those mixtures incorporating SRA and/or SAP.
2) SRA was the most effective shrinkage mitigation method compared to other
mitigation techniques.
3) PHCM achieved excellent early-age shrinkage reduction in balance with other
properties when used alone or combined with other shrinkage mitigation
techniques.
4) Combining PHCM and SRA mitigated the drawbacks of SRA including delays in
setting time and significant reduction in early-age compressive strength.
5) The addition of PHCM improved the shrinkage behaviour of mixtures
incorporating SAP and overcame the SAP drawbacks including higher mass loss
and porosity that usually lead to a reduction in early-age compressive strength.
6) Combining shrinkage mitigation techniques including SRA, SAP and PHCM
showed better overall behaviour compared to their individuals effects. This
includes enhanced compressive strength development and lower shrinkage.
Chapter 7 246
7.7. REFERENCES
Acker, P., (2004), “Swelling, shrinkage and creep: a mechanical approach to cement hydration,” Materials and Structures, Vol. 37, No. 4, pp. 237-243.
Aggoun, S., Cheikh-Zouaoui, M., Chikh, N. and Duval, R., (2008), “Effect of
some admixtures on the setting time and strength evolution of cement pastes at early ages,” Construction and Building Material, Vol. 22, No. 2, pp. 106-110.
Aitcin, P.C., (1999), “Demystifying autogenous shrinkage,” Concrete International, Vol.
21, No. 11, pp. 54-56. Anna, K., Leivo, M. and Sipari, P., (1995), “Experimental study on the basic phenomena
of shrinkage and cracking of fresh mortar,” Cement and Concrete Research, Vol. 25, No. 8, pp. 1747-1754.
Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafaï, N., (2006),
“Autogenous deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp. 123-136.
Beaudoin, J., Drame, H., Raki, L., and Alizadeh, R., (2008), “Formation and properties of
C-S-H-PEG nanostructures”, Materials and Structures, Vol. 42, No.7, pp. 1003-1014.
Bentz, D.P., (2005), “Capitalizing on self-desiccation for autogenous distribution of
chemical,” Proceedings of the 4th International Seminar on Self-Desiccation and Its Importance in Concrete Technology, B. Persson, D. Bentz, and L.O. Nilsson (eds.), Lund University, USA, pp. 189-196.
Bentz, D.P., (2006), “Influence of shrinkage-reducing admixtures on early-age properties
of cement pastes,” Journal of Advanced Concrete Technology, Vol. 4, No. 3, pp. 423-429.
Bentz, D.P., (2008), “A review of early-age properties of cement-based materials,”
Cement and Concrete Research, Vol. 38, No. 2, pp. 196-204. Bentz, D.P. and Jensen, O.M., (2004), “Mitigation strategies for autogenous shrinkage
cracking,” Cement and Concrete Composites, Vol. 26, No. 6, pp. 677-685. Bentz, D.P., Hansen, K.K., Madsen, H.D., Vallee, F. and Griese, E.J., (2001a),
“Drying/hydration in cement pastes during curing,” Materials and Structures, Vol. 34, No. 9, pp. 557-565.
Bentz, D.P., Geiker, M.R. and Hansen, K.K., (2001b), “Shrinkage-reducing admixtures
and early-age desiccation in cement pastes and mortars,” Cement and Concrete Research, Vol. 31, No. 7, pp. 1075-1085.
Chapter 7 247
Carde, C. and Francois, F., (1997), “Effect of the leaching of calcium hydroxide from cement paste on mechanical and physical properties,” Cement and Concrete Research, Vol. 27, No. 4, pp. 539-550.
Eberhardt, A.B. and Kaufmann, J., (2006), “Development of shrinkage reduced self-
compacting concrete,” ACI, Special Publication, SP-235, pp. 13-30. Hansen, T.C. and Nielsen, K.E.C., (1965), “Influence of aggregate properties on concrete
shrinkage,” ACI Materials Journal, Vol. 62, No. 7, 1965, pp. 783-794. He, Z., Li, Z.J., Chen, M.Z. and Liang, W.Q., (2006), “Properties of shrinkage-reducing
admixture-modified pastes and mortar,” Materials and Structures, Vol. 39, No. 4, pp. 445-453.
Hobbs, D.W., (1974), “Influence of aggregate restraint on the shrinkage of
concrete,” American Concrete Institution, Vol. 71, pp. 445-450. Holschemacher, K., Dehn, F., Klotz, S. and Weiße, D., (2005), “Experimental
investigation on ultra high-strength concrete under concentrated loading,” Proceedings of the Seventh International Symposium on Utilization of high-strength/ high performance concrete, Washington D.C., USA, Vol. 2, pp.1145-1158.
Ichinomiya, T., Hishiki, Y., Ohno, T., Morita, Y. and Takada, K., (2005), “Experimental
study on mechanical properties of ultra high-strength concrete with low-autogenous-shrinkage,” ACI, Special Publication, SP-228, pp. 1341-1352.
IUPAC, (1972). Manual of Symbols and Terminology, Appendix 2, Part 1. In: Colloid
and Surface Chemistry, Pure and Applied Chemistry, Vol. 31, 578 p. Jensen, O.M. and Hansen, P.F., (1996), “Autogenous deformation and change of relative
humidity in silica fume modified cement paste,” ACI Materials Journal, Vol. 93, No. 6, pp. 539-543.
Jensen, O.M. and Hansen, P.F., (2001), “Water-entrained cement-based materials: I.
Principles and theoretical background,” Cement and Concrete Research, Vol. 31, No. 4, pp. 647-654.
Jensen, O.M. and Hansen, P.F., (2002), “Water-entrained cement-based materials: II.
Experimental observations,” Cement and Concrete Research, Vol. 32, No. 6, pp. 973-978.
Kamen, A., Denarié, E., Sadouki, H. and Brühwiler, E., (2008), “Thermo-mechanical
response of UHPFRC at early-age: Experimental study and numerical simulation,” Cement and Concrete Research, V. 38, No. 6, pp. 822-831.
Chapter 7 248
Kovler, K., (1996), “Why Sealed Concrete Swells,” ACI Materials Journal, Vol. 93, No. 4, pp. 334-340.
Kovler, K. and Bentur, A., (2009), “Cracking sensitivity of normal and high strength
concretes,” ACI Materials Journal, Vol. 106, No. 6, pp. 537-542. Ma, J., Orgass, M., Dehn, F., Schmidt, D and Tue, N.V., (2004), “Comparative
investigations on ultra-high performance concrete with and without coarse aggregates,” Proceeding of the International Symposium on Ultra High Performance Concrete, Germany, pp. 205-212.
Matlese, C., Pistolesi, C., Lolli, A., Bravo. A., Cerulli, T. and Salvioni, D, (2005),
“Combined effect of expansion and shrinkage reducing admixture to obtain stable and durable mortars” Cement and Concrete Research, Vol. 35, No. 12, pp. 2244-2251.
Mikhail, R.S., Abo-El-Enein, S.A. and Gabr, N.A., (1977), “Hardened slag-cement pastes
of various porosities I. Compressive strength, degree of hydration and total porosity,” Journal of Applied Chemistry and Biotechnology, Vol. 24, No. 12, pp. 735-743.
Mönnig, S. and Lura, P., (2007), “Superabsorbent Polymers - An Additive to Increase the
Freeze-Thaw Resistance of High Strength Concrete,” In: Advances in Construction Materials, Part V, Springer, Berlin-Heidelberg, pp. 351-358.
Montgomery, D.C., (2009). Design and Analysis of Experiments, 7th Edition, John Wiley
& Sons, Hoboken, New Jersey, USA, 656 p. Odler, I. and Dörr, H., (1979), “Early hydration of tricalcium silicate I : Kinetics of the
hydration process and the stoichiometry of the hydration products,” Cement and Concrete Research, Vol. 9, No. 2, pp. 239-248.
Odler, I. and Rößler, M., (1985), “Investigations on the relationship between porosity,
structure and strength of hydrated Portland cement pastes: II. Effect of pore structure and of degree of hydration,” Cement and Concrete Research, Vol. 15, No. 3, pp. 401-410.
Powers, T.C., (1962), “A hypothesis on carbonation shrinkage,” Journal of Portland
Cement Association Research and Development Labs, Vol. 4, No. 2, 1962, pp. 40-50.
Rajabipour, F., Sant, G. and Weiss, J., (2008), “Interactions between shrinkage reducing
admixtures (SRA) and cement paste's pore solution,” Cement and Concrete Research, Vol. 38, No. 5, pp. 606-615.
Chapter 7 249
Ramachandran, V.S., Haber, R.M., Beaudoin, J.J. and Delgado, A.H., (2002). Handbook of Thermal Analysis of Construction Materials, Noyes Publications/William Andrew Publishing, Norwich, 655 p.
Sant, G., (2009), “Fundumental investigation related to the mitigation of volume change
in cement-based materials at early-ages,” Ph.D thesis, Purdue University, Indiana, USA, 203 p.
Tazawa, E. and Miyazawa, S., (1999), “Effect of constituents and curing condition on
autogenous shrinkage of concrete,” Proceedings of the International Workshop Autogenous Shrinkage of Concrete, (ed.) Eichi Tazawa, Taylor & Francis, pp. 269-280.
Wang, F., Zhou, Y., Peng, B., Liu, Z. and Hu, S., (2009), “Autogenous Shrinkage of
Concrete with Super-Absorbent Polymer,” ACI Materials Journal, Vol. 106, No. 2, 2009, pp. 123-127.
Weiss, J., Lura, P., Rajabipour, F. and Sant, G., (2008), “Performance of shrinkage-
reducing admixtures at different humidities and at early ages,” ACI Materials Journal, Vol. 105, No. 5, pp. 478-486.
Xiao, L. and Li, Z., (2008), “Early-age hydration of fresh concrete monitored by non-
contact electrical resistivity measurement,” Cement and Concrete Research, Vol. 38, No. 3, pp. 312-319.
Zhang, M. H., Tam, C. T. and Leow, M.P., (2003), “Effect of water-to-cementitious
materials ratio and silica fume on the autogenous shrinkage of concrete,” Cement and Concrete Research, Vol. 33, No. 10, pp. 1687-1694.
Chapter 8 250
CHAPTER EIGHT
INFLUENCE OF NATURAL WOLLASTONITE MICROFIBERS ON EARLY-AGE BEHAVIOUR OF UHPC*
In order to produce concrete characterized by lower early-age shrinkage and cracking risk
along with reducing its environmental and economic impact, the concept of reusing waste
concrete (i.e. partially hydrated cementitious materials) as a source for internal passive
restraining system and its efficiency was discussed in Chapters 6 and 7. In this chapter,
various sizes of natural wollastonite microfiber were added to the UHPC as a
replacement for cement. The potential of using wollastonite microfiber, as a natural
material, to improve early-age properties of UHPC along with achieving lower
environmental impact was investigated.
8.1. INTRODUCTION
Early-age cracking of cement based materials arises from the early rapid volume changes
as a result of autogenous and drying shrinkage and thermal deformations (Bentz et al.,
2008, Ma et al., 2007). Such volume changes induce tensile stresses within cement based
materials. The tensile strength of such materials and its ability to resist tensile stresses
increase with time (ACI Committee 231, 2010). Hence, a competition between the
induced tensile stresses and the development of cement based materials tensile strength
exists during early-ages. Once tensile stresses exceeded the tensile strength, micro-
*A version of this chapter has been submitted for review to the Journal of Materials in Civil Engineering, ASCE.
Chapter 8 251
cracking develops and propagates leading to visible shrinkage macro-cracking (ACI
Committee 231, 2010, Cusson, 2008). Shrinkage cracks later facilitate the penetration of
aggressive substances to concrete, leading to a reduction in its performance,
serviceability, and durability (Passuello et al., 2009).
Several methods have been advocated to minimize the cracking potential of
concrete, including using coarser cement particles, expansive additives, shrinkage
reducing admixtures and/or improving curing conditions (Bentz and Peltz, 2008, Van
Breugel and De Vries, 1998, Tazawa, 1998, Nmai et al., 1998). These approaches
primarily focus on reducing shrinkage strains in concrete, thereby reducing the level of
residual stress that develops (Shah et al., 1998). On the other hand, microfibers were
reported to act as a local restraint for shrinkage (Zhang and Li, 2001). Microfibers
generally bridge micro-cracks, leading to a reduction in crack widths and delaying the
occurrence of cracking (Lawler et al., 2003).
Different types of microfibers have been used as reinforcements for cementitious
materials, including organic, mineral, metallic, or synthetic microfibers (Andiç et al.,
2008, Pierre et al., 1997). Among these, metallic microfibers (e.g. steel) have been
widely used in different UHPC applications (Hoang et al., 2008). However, optimizing
the use of other types of microfibers in UHPC applications can provide an effective and
low-cost reinforcement method. A potential microfiber is wollastonite, which is widely
used in other industrial applications (e.g. ceramics, plastics, paints, etc.) (Azaroy et al.,
1995).
Chapter 8 252
Wollastonite is a naturally occurring, acicular, inert, white mineral (calcium meta
silicate [β- CaO-Si02]), which is less costly than steel and carbon microfibers (Low and
Beaudoin, 1992). Previous studies have shown the potential for using natural wollastonite
microfibers as a reinforcing material in cementitious materials (Low and Beaudoin, 1992,
Low and Beaudoin, 1993, Low and Beaudoin, 1994a). The addition of wollastonite in
cement-silica fume matrices showed significant improvements in pre-peak and post-peak
load, flexural toughness and ductility (Low and Beaudoin, 1993). Moreover, wollastonite
microfibers imbedded in cementitious materials achieved high stability without surface or
bulk deterioration with time (Low and Beaudoin, 1994b). However, there appears to be
little or no information with regards to the early-age properties, shrinkage and cracking
behaviour of UHPC reinforced with natural wollastonite microfibers. Moreover, the
mechanism of microfiber/matrix interfacial bond in UHPC has not yet been fully
characterized. Therefore, the aim of this study is to examine the feasibility of utilizing
various small-size wollastonite microfibers in UHPC to control shrinkage cracking, and
to investigate its effect on other early-age properties of UHPC.
8.2. RESEARCH SIGNIFICANCE
With the increasing use of UHPC in the construction, strengthening and rehabilitation of
different infrastructure, controlling the early-age shrinkage cracking of UHPC is essential
for ensuring an enhanced long-term performance and longer service life. Added
wollastonite microfibers acted as an internal restraint for shrinkage, reinforcing the
microstructure at the micro-crack level leading to an enhancement of early-age
engineering properties of UHPC matrix. From an environmental point of view, achieving
Chapter 8 253
UHPC with comparable performance, yet with lower cement content, using natural
wollastonite as partial replacement for cement can lead to a reduction in the cement factor
and consequently lower CO2 emissions. This can make the cement intensive UHPC more
sustainable.
8.3. EXPERIMENTAL PROGRAM
This experimental program aims to investigate the effect of wollastonite microfibers on
early-age behaviour of UHPC and their role in controlling shrinkage cracking. This is of
primary importance in order to achieve high performance and durable structures. In this
study, the hydration and strength development (compressive strength, flexural toughness,
heat of hydration, degree of hydration) and shrinkage behaviour have been investigated
on UHPC mixtures incorporating different contents and sizes of wollastonite microfibers.
All tests were conducted on UHPC specimens without heat curing in order to explore real
effects that govern UHPC shrinkage existing in structural elements cast in-situ.
8.3.1. Materials and Mixture Proportions
The materials used in this chapter were similar to that used in Chapter 3 (refer to Section
3.4.1). The chemical and physical properties of the used binders have been given in
Chapter 3 (Table 3-1). Commercially available natural wollastonite microfibers were
used at three dosages (4, 8 and 12%) as partial substitution for cement by volume. Three
microfiber sizes were used in this study: MF1 (length 152 µm, diameter 8 µm), MF2
(length 50 µm, diameter 5 µm), and MF3 (length 15 µm, diameter 3 µm) (see Fig. 8-1).
Chapter 8 254
The selected composition of the control mixture are shown in Chapter 3 (refer to Table
3-2). The characteristics of the tested mixtures are shown in Table 8-1.
(a) (b) (c)
Figure 8-1: Different sizes of wollastonite microfibers a) MF1, b) MF2 and c) MF3.
Table 8-1: Tested mixtures
Mixture MF1% MF2% MF3%
C ‐‐‐‐ ‐‐‐‐ ‐‐‐‐ M14 4.0 ‐‐‐‐ ‐‐‐‐ M18 8.0 ‐‐‐‐ ‐‐‐‐ M112 12.0 ‐‐‐‐ ‐‐‐‐ M24 ‐‐‐‐ 4.0 ‐‐‐‐ M28 ‐‐‐‐ 8.0 ‐‐‐‐ M212 ‐‐‐‐ 12.0 ‐‐‐‐ M34 ‐‐‐‐ ‐‐‐‐ 4.0 M38 ‐‐‐‐ ‐‐‐‐ 8.0 M312 ‐‐‐‐ ‐‐‐‐ 12.0
8.3.2. Preparation of Test Specimens and Testing Procedures
The experimental methods used in this chapter are the cubic compressive strength test,
flexural strength test, workability test, Semi-adiabatic calorimetry, TGA, MIP,
measurements of total free shrinkage, mass loss, setting time, shrinkage restraining test
Chapter 8 255
and SEM/EDX analysis. The cubic compressive strength test, Semi-adiabatic
calorimetry, TGA, MIP, measurements of total free shrinkage, mass loss, setting time and
SEM/EDX analysis were previously explained in chapters 3, 5 and 7. The workability,
flexural strength test and shrinkage restraining test are explained in the following
sections.
The workability of mixtures was evaluated based on the flow index (F), which is
defined as follows (Eq. 1):
100x0R
0R25RF(%)
−= Eq. 1
where R25 is the radius of the mortar pile after the 25th drop and R0 is the initial radius of
the mortar pile according to the ASTM C 1437 (Standard Test Method for Flow of
Hydraulic Cement Mortar).
Flexural strength testing was conducted on 35x35x200 mm UHPC prisms at the
ages of 3, 5 and 7 days (Pierre et al., 1997). The third-point loading flexural strength tests
were carried out using a computer controlled material testing system at a loading rate of
0.1 mm/min. Prior to the bending test, each beam specimen was maintained in a calcium
hydroxide saturated solution. A total of four specimens were tested for each mixture at
each age and the average result was reported.
An instrumented ring test was performed similar to previous work (See et al., 2003)
to quantify the restrained shrinkage behaviour, cracking age and width, for mixtures with
and without wollastonite microfibers. The dimensions of the concrete ring specimen used
in this evaluation are given in Fig. 8-2. The specimens were not allowed to dry from top
Chapter 8 256
and bottom surfaces of the ring by sealing the circumference. Each ring specimen was
equipped with four strain gages at the mid-height on the inner circumference of the steel
ring. Steel ring strain measurements were monitored until the concrete ring cracked. After
that, measurements of the cracking widths were taken every day for at least 7 days.
Figure 8-2: Concrete ring test.
8.4. RESULTS AND DISCUSSION
8.4.1. Workability
The workability of UHPC mixtures incorporating microfibers is highly affected by the
microfiber content and its aspect ratio (Tatnall, 2006). Figure 8-3 shows the relative
change in the flowability of UHPC mixtures incorporating different sizes and contents of
wollastonite microfibers with respect to that of the control mixtures. Incorporating MF1
and MF2 led to lower workability compared to that of mixtures without microfibers. The
higher the MF1 or MF2 content, the lower was the workability achieved. This can be
attributed to an increase in the microfibers interlocking due to its needle-like shape
(Tatnall, 2006). However, this reduction in workability can be overcome through
applying vibration during placing or using higher dosage of HRWRA (Tatnall, 2006,
Chapter 8 257
Ransinchung and Kumar, 2010). Conversely, adding MF3 increased workability of
UHPC mixtures compared to that of the control. The workability was better enhanced as
the added amount of MF3 increased. MF3 microfibers have a very small size; thus it
provides an internal lubricant effect by displacing water from voids between coarser
particles, leading to lower water demand (Nehdi et al., 1998, Schmidt, 1992).
-15
-10
-5
0
5
10
15
MF1 MF2 MF3Microfibers type
Rel
ativ
e ch
ange
in fl
owab
ility
(%)
4%8%12%
Figure 8-3: Relative change in flowability index for mixtures incorporating
different content and sizes of wollastonite microfibers compared to that of the
control mixture.
8.4.2. Early-Age Compressive Strength
Compressive strength is considered as a key property of UHPC. The addition of
wollastonite microfibers affected the early-age compressive strength of UHPC, as shown
in Fig. 8-4(a,b,c) and Table 8-2. The compressive strength development was mainly
influenced by wollastonite microfibers content and size (i.e. aspect ratio). During the very
early-age (i.e. ≤ 24 hrs), mixtures incorporating MF1 or MF2 exhibited compressive
Chapter 8 258
strength comparable to or higher than that of the control mixture (Fig. 8-4(a,b)). This
improvement in compressive strength was more significant at higher wollastonite
microfibers content. Conversely, the greater the MF3 content, the lower was the early-age
compressive strength (Table 8-2).
Table 8-2: Compressive strength results of UHPC mixtures incorporating different
content and sizes of wollastonite microfibers.
Microfibers Types MF1 (%) MF2 (%) MF3 (%)
Age (hrs)/Variation (%)* 4 8 12 4 8 12 4 8 12
12(hrs) 11.4 12.1 15.2 10.0 10.4 12.0 9.8 6.7 4.7 (%) (+21) (+28) (+61) (+6) (+10) (+27) (+4) (‐28) (‐50)
24(hrs) 37.6 38.0 39.1 36.9 37.5 38.1 36.4 33.0 29.3 (%) (+14) (+15) (+19) (+13) (+15) (+16) (+11) (0) (‐10)
168(hrs) 90.7 94.8 101.2 89.5 92.7 95.3 81.6 86.3 88.5 (%) (‐2) (+2) (+9) (‐3) (0) (+2) (‐12) (‐7) (‐4)
* Variation with respect to the control value at the same age **All compressive strength results are in (MPa)
The improvement in very early-age compressive strength can be ascribed to the
ability of microfibers to bridge micro-cracks, thus leading to a higher load carrying
capacity (Ding and Kusterle, 2000). The size-dependent effect of wollastonite microfibers
on early-age compressive strength can be explained as follows: The efficiency of
microfibers in bridging cracks is a function of the content and length of the used
microfibers and the fiber-paste interfacial bond strength (Hameed et al., 2009, Banthia
and Sheng, 1996). During very early-age, the degree of hydration is low and
consequently the microfiber/matrix bond strength is low (Chan and Li, 1997). Moreover,
the addition of wollastonite microfibers as a partial replacement for cement can delay the
bond strength development as a result of a dilution effect (Schmidt, 1992). Therefore,
Chapter 8 259
microfibers with a larger aspect ratio will be more effective as it has larger specific
surface area (i.e. contact area), thus resulting in sufficient anchorage length beyond the
edges of micro-crack. Moreover, the longer the hydration period, the higher was the
matrix strength and consequent improvement of the microfiber/matrix bond and bridging
efficiency, leading to higher compressive strength (Chan and Li, 1997).
0
2
4
6
8
10
12
14
16
0 5 10 15Wollastonite microfibers content (%)
Com
pres
sive
str
engt
h (M
Pa)
MF1-6hrsMF1-8hrsMF1-10hrsMF1-12hrs
(a)
Figure 8-4: Very early-age compressive strength development for different
wollastonite microfibers a) MF1, b) MF2 and c) MF3.
Chapter 8 260
0
2
4
6
8
10
12
14
16
0 5 10Wollastonite microfibers content (%)
Com
pres
sive
str
engt
h (M
Pa)
15
MF2-6 hrsMF2-8hrsMF2-10hrsMF2-12hrs
(b)
0
2
4
6
8
10
12
14
16
0 5 10Wollastonite microfibers content (%)
Com
pres
sive
str
engt
h (M
Pa)
15
MF3-6hrsMF3-8hrsMF3-10hrsMF3-12hrs
(C)
Figure 8-4 Contd’: Very early-age compressive strength development for different
wollastonite microfibers a) MF1, b) MF2 and c) MF3.
Chapter 8 261
At later ages (i.e. > 24 hrs), mixtures incorporating low wollastonite microfibers
content exhibited a slightly lower compressive strength compared to that of the control
mixture. However, increasing the wollastonite microfibers content diminished and/or
reduced the reduction in the compressive strength regardless of the microfiber size. For
instance, increasing the wollastonite microfiber content from 4 to 12% improved the 7-
days compressive strength from -2.0% to +9 % for MF1, -3% to +2% for MF2 and -12%
to -4% for MF3 of their respective control values, as shown in Fig. 8-5. A similar trend,
though with a smaller magnitude, was observed at 28-days. This can be attributed to an
improvement of the microfiber-matrix-bond with age, especially at the low w/c of UHPC,
which leads to higher microfiber crack-bridging efficiency (Chan and Li, 1997, Hamoush
et al., 2010). Moreover, adding very short microfibers (e.g. MF3) can enhance the
compressive strength due to its packing effect (Lange et al., 1997).
0.0
0.2
0.4
0.6
0.8
1.0
1.2
C M14 M18 M112 M24 M28 M212 M34 M38 M312
Mixtures
Rel
ativ
e st
reng
th
Figure 8-5: Relative early-age compressive strength of mixtures incorporating
different content and sizes of wollastonite microfibers at age 7 days.
Chapter 8 262
8.4.3. Heat of Hydration
A characteristic of cementitious materials is the heat generated due to the exothermic
hydration reactions of cement. This heat is translated into a temperature increase from
which the heat quantity developed can be evaluated (Chikh et al., 2008). Test results
were presented graphically in terms of semi-adiabatic temperature evolution versus time
in the present study.
Wollastonite microfiber is a relatively inert material expected not to have a
significant effect on the heat of hydration liberated during the hydration process (Low
and Beaudoin, 1992). Contradictory to this previous finding in (Low and Beaudoin,
1992), the addition of wollastonite microfibers in UHPC modified the evolution of heat
of hydration compared to that of the control mixture. Regardless of its size, adding 4% of
wollastonite microfibers exhibited higher heat of hydration with respect to that of the
control mixture (Fig. 8-6(a)). The higher the wollastonite microfibers content, the lower
was the increase in temperature peak with respect to that of the control mixture. For
instance, the temperature peak for incorporating 4 and 12% of wollastonite microfibers
varied by +13% and +3% for MF1, +11% and +1% for MF2 and by +6% and -5.5% for
MF3 of their respective control values (Fig. 8-6(b))
In the present study, very small size wollastonite microfibers were used and
different characterization of the hydration process was made compared to that in the
previous study (Low and Beaudoin, 1992). Therefore, the discrepancy in the heat of
hydration behaviour can be explained as follows: the added very fine inert particles (i.e.
MF3) displaced some of the water from voids between cement particles, making it
available; hence allowing more hydration reactions to take place (Schmidt, 1992).
Simultaneously, adding these inert materials as a partial replacement for cement induced
Chapter 8 263
a dilution effect, especially at higher dosages, leading to a lower hydration rate. UHPC is
characterized by a very low w/c. Hence, the availability of space for hydration products
to form is the main restriction on hydration development (Lea, 1988). Therefore, adding
4% MF3, which is finer than cement particles, displaced water from voids and led to a
higher rate of hydration. Conversely, increasing the MF3 content (> 4%), reduced the
active sites (i.e. dilution effect) and improved the packing density, thus limiting the
available space for hydration products to form and leading to a lower hydration rate.
On the other hand, the addition of 4% of MF1 and MF2 induced higher porosity
due to the percolation of microfibers (Bentz, 2000), which increased the space available
for hydration products. Consequently, more hydration reactions took place and a higher
temperature peak was reached. At higher contents, the dilution effect dominates the
hydration rate, resulting in a lower heat of hydration.
20
23
26
29
32
35
38
41
44
0 6 12 18 24 30 36 42 48Time (hours)
Tem
pera
ture
(°C
)
CM14M24M34
(a)
Figure 8-6: a) General trend of heat of hydration for wollastonite microfibers
mixtures and b) variation in temperature peak of wollastonite mixtures Vs. control
mixture.
Chapter 8 264
-15
-10
-5
0
5
10
15
MF1 MF2 MF3Microfibers
Vara
ition
in te
mp.
pea
k (%
)
4%
12% 8%
(b)
Figure 8-6 Contd’: a) General trend of heat of hydration for wollastonite
microfibers mixtures and b) variation in temperature peak of wollastonite mixtures
Vs. control mixture.
Figure 8-7(a,b) shows the degree of hydration and porosity for mixtures
incorporating different wollastonite microfiber sizes. It can be observed that, at the same
wollastonite microfibers content, the degree of hydration increased with higher size of
wollastonite microfibers, while the porosity did not change significantly. Hence, it is
expected that extra hydration products filled the additional space induced by the
incorporation of larger size of wollastonite microfibers. (Heat of hydration results for all
wollastonite microfibers can be found in Appendix D).
Chapter 8 265
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0 4 8 12 16 20 2Age(hours)
BW
(g/g
cem
ent)
4
CM18M28M38
(a)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
C M18 M28 M38Mixtures
Rel
ativ
e Po
rosi
ty
+5.4%
-7.8%+2.5%
(b)
Figure 8-7: a) Development of degree of hydration and b) relative porosity for
mixture incorporating 8% of wollastonite microfibers.
Chapter 8 266
8.4.4. Free Shrinkage and Mass Loss
The mass loss and total shrinkage curves for mixtures incorporating wollastonite
microfibers are shown in Fig. 8-8 and 8-9. The measured total shrinkage includes both
drying and autogenous shrinkage (thermal deformation can be ignored due to the small
cross-section of the specimens (Baroghel-Bouny et al., 2006)). For very low w/c (Tazawa
and Miyazawa, 1999), the contribution of autogenous shrinkage to the total shrinkage can
be comparable to that of drying shrinkage.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0 24 48 72 96 120 144 168Age (hours)
Mas
s lo
ss (%
)
CM14M24M34
(a)
Figure 8-8: Mass loss for mixtures incorporating wollastonite microfibers with a)
4%, b) 8% and c) 12% contents compared to that of the control mixture.
Chapter 8 267
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0 24 48 72 96 120 144 16Age (hours)
Mas
s lo
ss (%
)
8
CM18M28M38
(b)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0 24 48 72 96 120 144 168Age (hours)
Mas
s lo
ss (%
)
CM112M212M312
(c)
Figure 8-8 Contd’: Mass loss for mixtures incorporating wollastonite microfibers
with a) 4%, b) 8% and c) 12% contents compared to that of the control mixture.
Chapter 8 268
Despite the higher porosity induced by wollastonite microfibers, mass loss did not
change significantly compared to that of the control mixture with no microfibers.
Variation in mass loss due to incorporating different contents and sizes of wollastonite
microfibers, ranged between -3.5% to +4.6% with respect to that of the control mixture.
This can be attributed to the discontinuity of the pore structure induced by wollastonite
microfibers (Mathur et al., 2007). Moreover, this confirms previous results which
suggested that more hydration products were formed in the space induced by the
incorporation of wollastonite microfibers, thus leading to a denser and a discontinuous
pore structure.
The general trend of the total shrinkage results indicates a progressive reduction
in total shrinkage with increasing content and aspect ratio of wollastonite microfibers, as
shown in Fig. 8-9. As the microfibers content increased from 4% to 12%, the reduction in
total shrinkage at 7-days increased from 11% to 16% for MF1 and from 2% to 9% for
MF2 with respect to that of the control mixture. On the other hand, mixtures
incorporating different contents of MF3 exhibited a less significant reduction in the total
shrinkage compared to that of the control mixture.
Microfibers locally restrain shrinkage stresses in the cementitious matrix (Mangat
and Azari, 1984). As the matrix around shrinks, shear stresses develop along
microfibers, leading to compressive stresses in the microfibers and tensile stresses in the
cementitious matrix (Zhang and Li, 2001). The efficiency of microfibers in restraining
shrinkage depends on several parameters, including the elastic modulus of the
microfibers and matrix, the microfiber content and its aspect ratio (Zhang and Li, 2001,
Mangat and Azari, 1984). Hence, the shrinkage restraining efficiency of microfibers at
Chapter 8 269
early-age is higher due to the low elastic modulus of the cementitious matrix (Zhang and
Li, 2001). For instance, incorporating 4% of MF1 led to around 40% reduction in total
shrinkage at 24 hrs, and about 16% at 7-days compared to that of the control mixtures,
respectively. Moreover, for the same microfibers content, the higher the microfibers
aspect ratio, the longer was the effective length leading to higher microfiber-matrix
interfacial bond force and consequently higher restraint for shrinkage. For instance, at the
same microfiber content of 12%, incorporating MF1 (aspect ratio 19) reduced the total
shrinkage by about eight times greater than the reduction induced by MF3 (aspect ratio
5). Moreover, increasing the microfibers content led to better shrinkage reduction
(Mangat and Azari, 1984, Swamy and Stavrides, 1979). This can be attributed to the
smaller cross-sectional area of the matrix cylinder surrounding each microfiber needed to
induce adequate interfacial bond strength and shrinkage restraining (Zhang and Li, 2001).
-600
-500
-400
-300
-200
-100
00 24 48 72 96 120 144 16
Age (hours)
Tota
l shr
inka
ge (µ
ε)
8
CM14M24M34
(a)
Figure 8-9: a) General trend of total shrinkage for wollastonite microfibers mixtures and b) variation in total shrinkage of wollastonite mixtures Vs. control mixture at 7-
days.
Chapter 8 270
-20
-15
-10
-5
0
5
10
15
20MF1 MF2 MF3
Microfibers
Vara
ition
in to
tal s
hrin
kage
(%)
(b)
4%
12% 8%
Figure 8-9 Contd’: a) General trend of total shrinkage for wollastonite microfibers
mixtures and b) variation in total shrinkage of wollastonite mixtures Vs. control
mixture at 7-days.
Figure 8-10 shows the effect of adding wollastonite microfibers on autogenous
shrinkage development. Generally, mixtures incorporating wollastonite microfibers
exhibited lower autogenous shrinkage compared to that of the control mixture. In
addition, the MF3 mixture achieved the lowest autogenous shrinkage compared to that of
other wollastonite microfiber mixtures. Autogenous shrinkage is strongly related to
hydration reactions in which water is consumed, leading to internal self-desiccation
(without external water loss) (Powers and Brownyard, 1948). Adding wollastonite
microfibers as a partial replacement for cement led to a lower hydration activity (i.e.
dilution effect) (Bentz et al., 2008), and locally restrained shrinkage as shown earlier.
Hence, increasing the wollastonite microfibers content increased the passive internal
restraint along with reducing hydration activity. Consequently, lower autogenous
Chapter 8 271
shrinkage occurred (Bentz et al., 2008, Bentz and Peltz, 2008). Moreover, mixtures
incorporating the very fine microfibers (i.e. MF3) exhibited early age expansion, which
led to a lower net autogenous shrinkage. This expansion can be attributed to the
development of disjoining pressure as a result of water absorption on these very fine
microfiber surfaces (Craeye et al., 2010). (Shrinkage results for all wollastonite
microfibers can be found in Appendix D).
-600
-500
-400
-300
-200
-100
0
1000 24 48 72 96 120 144 1
Age (hours)
Aut
ogen
ous
stra
in (µ
ε)
68
CM14M24M34
(a)
Figure 8-10: a) General trend of autogenous shrinkage for wollastonite microfibers
mixtures and b) variation in autogenous shrinkage of wollastonite mixtures Vs.
control mixture at 7-days.
Chapter 8 272
-40
-30
-20
-10
0
10
20
30
40MF1 MF2 MF3
Microfibers
Vara
ition
in a
utog
enou
s sh
rinka
ge
(%)
(b)
4%
12% 8%
Figure 8-10 Contd’: a) General trend of autogenous shrinkage for wollastonite
microfibers mixtures and b) variation in autogenous shrinkage of wollastonite
mixtures Vs. control mixture at 7-days.
8.4.5. Restrained Shrinkage
MF1 and MF2 wollastonite microfibers were found to enhance the shrinkage cracking
resistance as it delayed the age at which the first crack occurred, compared to that of the
control mixture. For instance, adding 4% of MF1 and MF2 delayed the cracking age by
about 27% and 16% compared to that of the control mixture, respectively. The higher the
MF1 or MF2 content, the later was the onset of cracking (Fig. 8-11(a)). Conversely,
increasing the MF3 content reduced the cracking age (Fig. 8-11(b)). In the absence of
microfibers, there is little resistance to crack propagation, while in a cementitious matrix
incorporating microfibers, the weakest crack resistance point changes with time
depending on the distribution of microfibers (Passuello et al., 2009). Hence, microfibers
Chapter 8 273
delayed the coalescence and propagation of cracks at early-age through better stress
transfer at micro-cracks (Banthia and Sheng, 1996, Swamy and Stavrides, 1979). Using
microfibers with a very small aspect ratio (i.e. MF3) jeopardized its crack bridging
efficiency, leading to lower shrinkage cracking resistance and earlier crack formation. On
the other hand, the increase in the cracking age with increasing microfiber content
implies a higher crack bridging efficiency and ability of larger size wollastonite
microfibers to overcome the reduction in matrix strength induced by dilution effect.
0
10
20
30
40
50
60
0 20 40 60 80 100 120 140 16Time (hours)
Stra
in(µε)
0
CM14M18M112
(a)
Figure 8-11: Strain measurements of the steel ring for mixtures incorporating different content of a) MF1 and b) MF3.
Chapter 8 274
0
10
20
30
40
50
60
0 20 40 60 80 100 120 140 160Time (hours)
Stra
in(µε)
CM34M38M312
(b)
Figure 8-11 Contd’: Strain measurements of the steel ring for mixtures incorporating different content of a) MF1 and b) MF3.
For all tested mixtures, cracking was first observed to occur after approximately
42 hours. Steel rings containing specimens with MF1 and MF2 microfibers tended to
experience the development of two cracks, while those with MF3 or no microfibers
exhibited only one large crack. The presence of multiple cracks is an expected result,
since microfibers create internally several small restraints inside the matrix leading to
lower crack width as the cracking opening is shared by all cracks (Passuello et al., 2009).
Moreover, the total width of the multiple cracks developed in MF1 and MF2 was
lower than that of the control mixture single crack at the end of the investigated period
(Fig. 8-12). The absence of multiple cracks in the control and MF3 mixtures can be
attributed to the stress relief induced by the opening of single crack, since, unlike
matrices with microfibers, there is no ability to transfer or redistribute stresses (Shah and
Chapter 8 275
Weiss, 2006). Moreover, the crack width increased with increasing MF3 content. This
can be ascribed to a reduction in the tensile strength of the matrix due to the cement
dilution effect.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
C M14 M18 M112 M24 M28 M212 M34 M38 M312Mixtures
Tota
l cra
ck w
idth
(mm
)
Figure 8-12: Total crack width for mixtures incorporating different content and
sizes of wollastonite microfibers.
8.4.6. Flexural Strength
The addition of wollastonite microfibers in UHPC mixture significantly modified its
flexural strength characteristics compared to that of the control mixture without
microfibers. The flexural strength varied depending on the wollastonite microfibers
content and aspect ratio and the hydration period of the cementitious matrix.
Figure 8-13 shows the variation in flexural strength for different wollastonite
microfibers contents and aspect ratios relatively to that of the control mixture. At 3-days,
the flexural strength increased as the content of the wollastonite microfibers and its
Chapter 8 276
aspect ratio increased, except for MF3. Moreover, at 5 and 7-days, tested specimens of
mixtures incorporating MF1 and MF2 did not display signs of distress during the initial
portion of the test procedure. Popping and cracking sounds could be heard towards the
end of the test as the load applied to the specimen continued to increase. Subsequently, a
sudden and brittle failure similar to that of the control mixture without microfibers
occurred.
-30
-20
-10
0
10
20
30
M14 M18 M112 M24 M28 M212 M34 M38 M312Mixtures
Varia
tion
in fl
exur
al s
tren
gth
(%) 3 days
5 days7 days
Figure 8-13: Relative flexural strength for mixtures incorporating different content
and sizes of wollastonite microfibers compared to that of the control mixture.
The cracking process is believed to initiate as sub-micro cracks formed around
flaws once the load is applied. Coalescence of these sub-micro cracks results in forming
narrow micro-cracks. Microfibers hindered the widening of these narrow micro-cracks,
though their lengths continued to increase. Eventually, the coalescence of narrow micro-
Chapter 8 277
cracks localizes deformation, producing macro-cracks over the entire width of specimens
(Lawler et al., 2003). Therefore, increasing the amount of microfibers increases the crack
bridging efficiency (Lawler et al., 2003, Banthia and Sheng, 1996) and consequently
delays the development of macro-cracks, which leads to a stronger material as observed
for the mixtures incorporating MF1 and MF2 (Lawler et al., 2003). The
microfiber/matrix interfacial bond and fiber length are important factors that can
influence the effectiveness of microfibers as a medium of stress transfer (Banthia and
Sheng, 1996). Hence, increasing the embedment length and the fiber perimeter increases
the interfacial contact area between the fibers and matrix, leading to stronger interfacial
bond (Pakravan et al., 2010). This can explain the increase in flexural strength as the
aspect ratio of the added wollastonite microfibers increased. On the other hand, it seems
that the very small size of MF3 and the expected reduction in the strength induced by the
cement dilution effect significantly reduced the flexural strength of MF3 mixtures.
The very high strength of the UHPC matrix is expected to enhance the fiber-matrix
interfacial bond, providing high resistance to fiber pullout. Previous studies (Ransinchung
and Kumar, 2010, Low and Beaudoin, 1994) showed that wollastonite microfibers have
the ability to react with cement and develop a strong chemical bond. This strong chemical
adhesion will increase the bond strength, leading to fiber rupture rather than pull-out
(Rathod and Patodi, 2010). Hence, the reduction in the flexural strength, at ages 5 and 7
days, can be attributed to sudden rupture of wollastonite microfibers under flexural load
(Rathod and Patodi, 2010, Lawler et al., 2005). This was confirmed by SEM micrographs
as shown in Fig. 8-14. Moreover, wollastonite microfibers have a relatively low modulus
of elasticity (30 GPa) (Pierre et al., 1997) while the UHPC matrix has a very high
Chapter 8 278
modulus (around 34 GPa after 3 days). This reduced the apparent strength of inclined
microfibers (Zhang and Li, 2002). Thus, the amount of microfibers that pulled out instead
of rupturing was reduced and the fracture toughness decreased.
Figure 8-14: Scanning electron micrograph and energy dispersive X-ray element
analysis for ruptured wollastonite microfibers.
8.6. CONCLUSIONS
The feasibility of utilizing various small-size wollastonite microfibers in UHPC to
control shrinkage cracking and its effect on other early-age properties of UHPC were
investigated. The main conclusions that can be drawn from this experimental are the
following:
1) Influence of wollastonite microfibers addition on UHPC mixtures early-age
compressive strength is highly influence by its bridging micro-cracks efficiency,
packing and dilution effects.
Chapter 8 279
2) Addition of high aspect ratio of wollastonite microfibers can improve UHPC
hydration process through providing more space for hydration products to form.
3) Addition of wollastonite microfibers appears to promote pore discontinuity
leading to lower mass loss and drying shrinkage.
4) Addition of wollastonite microfibers can act as passive internal restraint leading
to lower measured shrinkage.
5) The low UHPC matrix elastic modulus at early-age increases the shrinkage
restraining effect of wollastonite microfibers.
6) Addition of wollastonite microfibers retarded development of shrinkage cracks as
it delay the coalescence of micro-cracks.
7) Addition of wollastonite microfibers increase the pre-peak load, however, no
improvement in ductility was achieved due to rupture of wollastonite microfibers.
Chapter 8 280
8.7. REFERENCES
ACI Committee 231, (2010). Report on Early-Age Cracking: Causes, Measurement, and Mitigation, Technical committee document No. 231.R-10, American Concrete Institute, Farmington Hills, 48 p.
Andiç, Ö., Yardımcı, M.Y. and Ramyar, V., (2008), “Performance of carbon,
polyvinylalcohol and steel based microfibers on alkali-silica reaction expansion,” Construction and Building Materials, Vol. 22, No. 7, pp. 1527-1531.
Azarov, G.M., Maiorova, E.V., Oborina, M.A. and Belyakov, A.V., (1995),
“Wollastonite raw materials and their applications (a review),” Glass and Ceramics, Vol. 52, No. 9, pp. 237-240.
Banthia, N. and Sheng, J., (1996), “Fracture toughness of microfiber reinforced cement
composites,” Cement and Concrete Composites, Vol. 18, No. 4, pp. 251-269. Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafai, N., (2006)
“Autogenous deformations of cement pastes: Part II. w/c effects, micro-macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp.123-136.
Bentz, D.P., (2000), “Fibers, percolation, and spalling of high performance concrete,”
ACI Materials Journal, Vol. 97, No. 3, pp. 351-359. Bentz, D.P., and Peltz, M.A., (2008), "Reducing thermal and autogenous shrinkage
contributions to early-age cracking," ACI Materials Journal, Vol. 105, No. 4, pp. 414-420.
Bentz, D.P., Sant, G. and Weiss, J., (2008), “ Early-age properties of cement-based
materials I: influence of cement fineness,” Journal of Materials in Civil Engineering, Vol. 20, No. 7, pp. 502-508.
Chan, Y.W. and Li, V.C., (1997), “Age effect on the characteristics of fiber/cement
interfacial properties,” Journal of Materials Science, Vol. 32, No. 19, pp. 5287-5292.
Chikh, N., Cheikh-Zouaoui, M., Aggoun, S. and Duval, R., (2008), “Effects of calcium
nitrate and triisopropanolamine on the setting and strength evolution of Portland cement pastes,” Materials and Structures, Vol. 41, No. 1, pp. 31-36.
Craeye, B., De Schutter, G., Desmet, B., Vantomme, J., Heirman, G., Vandewalle, L.,
Cizer, Ö., Aggoun, S. and Kadri, E.H., (2010), “Effect of mineral filler type on autogenous shrinkage of self-compacting concrete,” Cement and Concrete Research, Vol. 40, No. 6, pp. 908-913.
Chapter 8 281
Cusson, D., (2008), “Effect of blended cements on effectiveness of internal curing of HPC,” Proceedings of the Internal Curing of High-Performance Concretes: Laboratory and Field Experiences, ACI SP-256, American Concrete Institute, Farmington Hills, pp. 105-120.
Ding, Y. and Kusterle, W., (2000), “Compressive stress-strain relationship of steel fiber-
reinforced concrete at early-age,” Cement and Concrete Research, Vol. 30, No. 10, pp. 1573-1579.
Hameed, R., Turatsinze, A., Duprat, F. and Sellier, A., (2009), “Metallic fiber reinforced
concrete : effect of fiber aspect ratio on the flexural properties,” ARPN Journal of Engineering and Applied Sciences, Vol. 4, No. 5, pp. 67-72.
Hamoush, S., Abu-Lebdeh, T., Cummins, T. and Zornig, B., (2010), “Pullout
characterizations of various steel fibers embedded in very high-strength concrete,” American Journal of Engineering and Applied Sciences, Vol. 3, No. 2, pp. 418-426.
Hoang, K.H., Phat, H.B., Hien, L.D. and Chanh, N.V., (2008), “Influence of types of
steel fiber on properties of ultra high performance concrete,” Proceedings of the third ACF International Conference: Sustainable Concrete Technology and Structures in Local Climate and Environment conditions, Ho Chi Minh city, Vietnam, pp. 347-355.
Lange, F., Mörtel, H. and Rudert, V., (1997), “Dense packing of cement pastes and
resulting consequences on mortar properties,” Cement and Concrete Research, Vol. 27, No. 10, pp. 1481-1488.
Lawler, J.S., Wilhelm, T., Zampini, D. and Shah, S.P., (2003), “Fracture processes of
hybrid fiber-reinforced mortar,” Materials and Structures, Vol. 36, No. 3, pp. 197-208.
Lawler, J.S., Zampini, D. and Shah, S.P., (2005), “Microfiber and macrofiber hybrid
fiber-reinforced concrete,” Journal of Materials in Civil Engineering, Vol. 17, No. 5, pp. 595-604.
Lea, F.M., (1988). Lea's Chemistry of Cement and Concrete, 4th Edition, (ed.) P.C.
Hewlett, J. Wiley, New York, 1053 p. Low, N.M.P. and Beaudoin, J.J., (1992), ‘Mechanical properties of high performance
cement binders reinforced with wollastonite microfibers,” Cement and Concrete Research, Vol. 22, No. 5, pp. 981-989.
Low, N.M.P. and Beaudoin, J.J., (1993), “Flexural strength and microstructure of cement
binders reinforced with wollastonite microfibers,” Cement and Concrete Research, Vol. 23, No. 4, pp. 905-916.
Chapter 8 282
Low, N.M.P. and Beaudoin, J.J., (1994a), “The flexural toughness and ductility of portland cement-based binders reinforced with wollastonite microfibers,” Cement and Concrete Research, Vol. 24, No. 2, pp. 250-258.
Low, N.M.P. and Beaudoin, J.J., (1994b), “Stability of portland cement-based binders
reinforced with natural wollastonite microfibers,” Cement and Concrete Research, Vol. 24, No. 5, pp. 874-884.
Ma, B., Wang, X., Liang, W., Li, X. and He, Z., (2007), “Study on early-age
cracking of cement-based materials with superplasticizer,” Construction and Building Materials, Vol. 21, No. 11, pp. 2017-2022.
Mangat, P.S. and Azari, M.M., (1984), “A theory for the free shrinkage of steel fiber
reinforced cement matrices,” Journal of Materials Science, Vol. 19, No. 7, pp. 2183-2194.
Mathur, R., Misra, A.K. and Goel, P., (2007), “Influence of wollastonite on mechanical
properties of concrete,” Journal of Scientific and Industrial Research, Vol. 66, No. 12, pp.1029-1034.
Nehdi, M., Mindess, S. and Aïtcin, P.-C., (1998), “Rheology of high-performance
concrete: effect of ultrafine particles,” Cement and Concrete Research, Vol. 28, No. 5, pp. 687-697.
Nmai, C., Tomita, R., Hondo, F. and Buffenbarger, J., (1998), “Shrinkage reducing
admixtures,” Concrete International, Vol. 20, No. 4, pp. 31-37. Pakravan, H. R., Jamshidi, M. and Latifi, M., (2010), “Performance of fibers embedded
in a cementitious matrix,” Journal of Applied Polymer Science, Vol. 116, No. 3, pp. 1247-1253.
Passuello, A., Moriconi, G. and Shah, S.P., (2009), “Cracking
behavior of concrete with shrinkage reducing admixtures and PVA fibers,” Cement and Concrete Composites, Vol. 31, No. 10, pp. 699-704.
Pierre, P., Pleau, R., Rhéaume, M. and Pigeon, M., (1997), “The influence of microfiber
on the drying behaviour of cement paste and mortars,” Proceedings of Annual conference of the Canadian society for civil engineering, Sherbrooke, Québec, pp. 123-131.
Powers, T.C. and Brownyard, T.L., (1948), “Studies of the physical properties of
hardened portland cement paste,” Bulletin No. 22., Research Laboratories of the Portland Cement Association. Reprinted from Journal of the American Concrete Institute, October 1946–April 1947, Proceedings 43, Detroit, pp. 971-992.
Chapter 8 283
Ransinchung, G.D., and Kumar, B., (2010), “Investigations on pastes and mortars of ordinary portland cement admixed with wollastonite and microsilica,” Journal of Materials in Civil Engineering, Vol. 22, No. 4, pp. 305-313.
Rathod, J.D. and Patodi, S.C., (2010), “Interface tailoring of polyester-type fiber in
engineered cementitious composite matrix against pullout,” ACI Materials Journal, Vol. 107, No. 2, pp. 114-122.
Schmidt, M., (1992), “Cement with inter-ground additives: capabilities and
environmental relief (I),” Zement Kalk Gips, Vol. 45, No. 7, pp. 64-69. See, H.T., Attiogbe, E.K. and Miltenberger, M.A., (2003), “Shrinkage cracking
characteristics of concrete using ring specimens,” ACI Materials Journal, Vol. 100, No. 3, pp. 239-245.
Shah, S.P., Weiss, W.J. and Yang, W., (1998), “Shrinkage cracking-can it be
prevented?,” Concrete International, Vol. 20, No. 4, pp. 51-55. Shah, H. R. and Weiss, J., (2006), “Quantifying shrinkage cracking in fiber reinforced
concrete using the ring test,” Materials and Structures, Vol. 39, No.9, pp. 887-899.
Swamy, R.N. and Stavrides, H., (1979), “Influence of fiber reinforcement on restraining
shrinkage and cracking,” ACI Journal, Vol. 75, No. 3, pp. 443-460. Tatnall, P.C. (2006), “Chapter 49: Fiber-Reinforced Concrete,” In Significance of Tests
and Properties of Concrete and Concrete-Making Materials (STP 169D), Lamond, J. F. and Pielert, J. H. (eds.), ASTM International, West Conshohcken, PA., pp. 578-594.
Tazawa, E., (1998). Autogenous Shrinkage of Concrete. Proceedings of the International
Workshop on Autogenous Shrinkage of Concrete, Japan Concrete Institute, Hiroshima, Japan, E & FN Spon, New York, 358 p.
Tazawa, E. and Miyazawa, S., (1999), “Effect of constituents and curing condition on
autogenous shrinkage of concrete,” Proceedings of the International Workshop Autogenous Shrinkage of Concrete, Eichi Tazawa(ed.), Taylor & Francis, pp. 269-280.
Van Breugel, K. and De Vries, J., (1998), “Mixture optimization of low water/cement
ratio high strength concretes in view of autogenous shrinkage,” Proceedings of International Symposium on High Performance and Reactive Powder Concrete, Sherbrooke, Canada, pp. 365-382.
Chapter 8 284
Zhang, J., and Li, V., (2001), “Influences of fibers on drying shrinkage of fiber-reinforced cementitious composite,” Journal of Engineering Mechanics, Vol. 127, No. 1, pp. 37-44.
Zhang, J. and Li, V., (2002), “Effect of inclination angle on fiber rupture load in fiber
reinforced cementitious composites,” Composite Science and technology, Vol. 62,
No. 6, pp. 775-781.
Chapter 9 285
CHAPTER NINE
SHRINKAGE BEHAVIOUR OF UHPC WITH SHRINKAGE REDUCING ADMIXTURE AND WOLLASTONITE
MICROFIBERS*
In previous chapters, shrinkage reducing admixtures and wollastonite microfibers have
been used to improve shrinkage behaviour and the cracking potential of UHPC.
Shrinkage reducing admixture had shown a high efficiency in reducing early-age
shrinkage, however, it delays development of mechanical properties. Wollastonite
microfibers was found to act as an internal passive restraint for shrinkage, reinforcing the
microstructure at the micro-crack level leading to an enhancement of early-age
engineering properties of UHPC matrix. These results are based on previous
independently analysis of SRA and wollastonite microfibers, without considering their
combination. In this chapter, the improvement in early-age properties, shrinkage
behaviour and reduction of cracking potential of UHPC due to the addition of the SRA
and wollastonite microfibers, used separately or in a blend, were investigated and
analyzed.
9.1. INTRODUCTION
General background on the UHPC properties (high strength and enhanced durability) and
characterization (homogeneity, stronger and higher packing density microstructure and its
very low water to cement ratio) has been given in previous chapter 3. In addition, it has
*A version of this chapter has been submitted for review to Cement and Concrete Composites.
Chapter 9 286
been illustrated that UHPC has high tendency to early-age cracking which probably affect
its durability. Furthermore, UHPC has a serious environmental and economical impact
due to its high cement content which increases energy consumption and CO2 emissions.
In Chapter 8, it was outlined the potential of replacing cement with an inert
natural material, such as wollastonite microfibers. Addition of wollastonite microfibers
was found to improve the early-age properties of UHPC. This probably will lead to
environmental and economic benefits.
Among the several methods that have been proposed to minimize cracking
potential of concrete, shrinkage reducing admixture (SRA) was the most commonly used
technique. SRAs directly influence shrinkage by decreasing the surface tension of the
pore solution, leading to lower capillary stresses and consequently a reduced shrinkage
(Nmai et al., 1998). However, possible disadvantages of these materials include
effectiveness reduction over time (Schemmel et al., 1999), delaying setting time,
reducing strength (Bentz, 2006) and washing out of SRA as illustrated in chapter 5.
Therefore, the aim of this research is to develop a strategy for producing UHPC
with lower cracking propensity to traditional UHPC, while having a more positive
environmental foot-print, through combining the benefits of SRA and wollastonite
microfibers.
9.2. RESEARCH SIGNIFICANCE
Reducing concrete shrinkage cracks is important for durability, as well as that of strength,
particularly in structures such as slabs, bridge deck overlays, tunnel linings and
Chapter 9 287
pavements. In this chapter, the reduction of cracking risk has been evaluated by the
reduction of shrinkage and the increase of the crack opening resistance due to the
addition of shrinkage reducing admixtures (SRA) and wollastonite microfibers. Both
technologies have been considered individually and used in combination. Addition of
wollastonite microfibers had jeopardize the reduction It has been noted that combining
SRA and wollastonite microfibers has led to a better cracking behaviour compared to
mixtures incorporating one of these techniques alone. Moreover, the combining effect
gives better carking behaviour than the effect of increasing SRA or wollastonite
microfibers dosages.
9.3. EXPERIMENTAL PROGRAM
This experimental program aims to investigate the effect of SRA and wollastonite
microfibers on early-age behaviour of UHPC and their role in controlling shrinkage
cracking. This is of primary importance in order to achieve high performance and durable
structures. In this study, monitoring of the strength development (compressive strength,
heat of hydration, degree of hydration) and characterization of the shrinkage behaviour
have been investigated on UHPC mixtures incorporating different dosages of SRA and/or
wollastonite microfibers. All tests were conducted on UHPC specimens without heat
curing in order to understand real effects that govern UHPC shrinkage in structural
elements cast in situ.
9.3.1. Materials and Mixture Proportions
The materials used in this chapter were similar to that used in Chapter 3 (refer to Section
3.4.1). The chemical and physical properties of the used binders have been given in
Chapter 9 288
Chapter 3 (refer to Table 3-1). A poly-oxyalkylene alkyl ether shrinkage reducing
admixture was used in this study. SRA dosages of 1% and 2% by mass of cement were
added as a partial replacement of the mixing water. Commercially available natural
wollastonite microfibers MF1 (length 152 µm, diameter 8 µm) were used at two
concentrations (4 and 12%) as partial substitution for cement by volume. The selected
composition of the control mixture are shown in Chapter 3 (refer to Table 3-2). The
characteristics of the tested mixtures are shown in Table 9-1.
Table 9-1: Tested mixtures.
Mixture SRA (%) Wollastonite microfiber (%)
C ‐‐‐‐ ‐‐‐‐ R1 1.0 ‐‐‐‐ R2 2.0 ‐‐‐‐ W4 ‐‐‐‐ 4.0 W12 ‐‐‐‐ 12.0 R1W4 1.0 4.0 R1W12 1.0 12.0 R2W1 2.0 4.0 R2W12 2.0 12.0
9.3.2. Preparation of Test Specimens and Testing Procedures
The experimental methods used in this chapter are the cubic compressive strength test,
Semi-adiabatic calorimetry, TGA, MIP, measurements of total free shrinkage, mass loss,
setting time, shrinkage restraining test, COD and SEM/EDX analysis. All tests were
previously explained in chapters 3, 5, 7 and 8.
Chapter 9 289
9.4. RESULTS AND DISCUSSION
9.4.1. Compressive Strength
Compressive strength is considered as a key property of UHPC. The addition of
wollastonite microfibers was found to affect the early-age compressive strength of
UHPC, as shown in Fig. 9-1. A detailed discussion on the effect of wollastonite
microfibers addition on compressive strength was given in Chapter 8.
On the other hand, mixtures incorporating SRA achieved a lower early-age
compressive strength compared to that of the control mixture (Fig. 9-1). The higher the
SRA dosage, the lower was the achieved early-age compressive strength. This can be
attributed to the retardation effect induced by SRA, in agreement with previous work (He
et al., 2006).
0
5
10
15
20
25
30
35
40
45
6 8 10 12 24Age (hours)
Com
pres
sive
str
engt
h (M
Pa) C
W4W12R1R2
Figure 9-1: Development of early-age compressive strength of UHPC mixtures
with and without different dosages of SRA or wollastonite microfibers.
Chapter 9 290
Figure 9-2 illustrates the combined effect of SRA and wollastonite microfibers on
the very early-age compressive strength. Generally, adding wollastonite microfibers to
SRA mixtures improved the early-age compressive strength compared to that of mixtures
incorporating SRA alone. The higher the wollastonite microfibers content, the higher was
the improvement in compressive strength. For instance, adding 4% and 12% of
wollastonite microfibers was found to increase the early-age compressive strength of
mixture R1 by 12% and 20%, respectively and by 55% and 70% at age 24 hrs,
respectively for mixture R2 (Fig. 9-2(a)). Compared to mixtures incorporating
wollastonite microfibers alone, mixtures incorporating both SRA and wollastonite
microfibers exhibited lower early-age compressive strength. The higher the added dosage
of SRA, the higher was the reduction in the compressive strength. For instance, adding 1
and 2% of SRA was found to reduce the 24 hrs compressive strength for mixture W4 by
6% and 16%, respectively and by 4% and 13% at age 24 hrs, respectively for mixture
W12 (Fig. 9-2(b)).
The combined effect of SRA and wollastonite microfibers on the very early-age
compressive strength represents the net result of two conflicting effects: reinforcing and
retardation. Wollastonite microfibers bridge micro-cracks and reinforce the
microstructure, leading to higher strength (compared to that of the R1 and R2 mixtures).
Concurrently, SRA retards the hydration process and consequently reduces the
reinforcing efficiency of wollastonite microfibers (compared to that the W4 and W12
mixtures). The efficiency of microfibers in bridging micro-cracks is a function of the
interfacial microfiber/matrix bond strength (Hameed et al., 2009, Banthia and Sheng,
1996). The lower the degree of hydration, the smaller is the microfiber/matrix bond
Chapter 9 291
strength, and consequently the lower is the reinforcing efficiency (Chan and Li, 1997).
This is shown in degree of hydration results discussed below.
0
5
10
15
20
25
30
35
40
45
6 8 10 12 24Age (hours)
Com
pres
sive
str
engt
h (M
Pa) C
R1R1W4R1W12R2R2W4R2W12
(a)
0
5
10
15
20
25
30
35
40
45
6 8 10 12 24Age (hours)
Com
pres
sive
str
engt
h (M
Pa) C
W4R1W4R2W4W12R1W12R2W12
(b)
Figure 9-2: Effect of combing SRA and wollastonite microfibers on early-age
compressive strength of UHPC with respect to a) SRA and b) wollastonite
microfibers mixtures.
Chapter 9 292
At later ages (i.e. >24 hrs), similar to very early-age results, the addition of
wollastonite microfibers enhanced the compressive strength at late ages for SRA
mixtures, as shown in Fig. 9-3. The micro-cracks bridging and reinforcing of the
microstructure induced by wollastonite microfibers was able to reduce the negative effect
induced by SRA and achieve acceptable compressive strength behaviour.
-15
-10
-5
0
5
10
15
7 Days 28 Days
Com
pres
sive
str
engt
h di
ffern
ce (%
)
W4 W12R1 R2R1W4 R1W12R2W4 R2W12
Figure 9-3: Variation in the compressive strength of UHPC mixture with different
dosages of SRA and/or wollastonite microfibers compared to the control mixture.
9.4.2. Heat of Hydration
In chapter 8, the wollastonite microfibers effect on heat hydration was investigated and
analyzed. It was concluded that addition wollastonite microfibers can alert the hydration
process through inducing higher porosity, due to percolation of microfibers (Bentz, 2000)
Chapter 9 293
and displacing water from pores. Adding wollastonite microfiber exhibited slightly
higher heat of hydration with respect to that of control mixture (Fig. 9-4(a)).
It can be noted that adding SRA retarded hydration reactions and reduced the
liberated heat compared to that of the control mixture without SRA, as shown in Fig. 9-
4(a). This is consistent with previous results (Eberhardt and Kaufmann, 2006).
Furthermore, as the SRA dosage increased from 1 to 2%, the temperature peak decreased
and occurred later in time compared to that of the 1% SRA mixture. This retardation
effect can be attributed to a reduction in the ability of salts (e.g. alkali sulphates) to
dissolve and ionize in the pore solution due to the lower polarity induced by SRA
addition (Rajabipour et al., 2008).
On the other hand, mixtures incorporating both wollastonite microfibers and SRA
exhibited similar trend to that of mixtures incorporating SRA alone, regardless of the
added amount of wollastonite microfibers, as shown in Fig. 9-4(b). The higher the dosage
of SRA, the smaller and later the temperature peaks occurred. Moreover, at the same
dosage of SRA, on significant change in the heat of hydration was found as the content of
wollastonite microfibers increased from 0% to 12%. For instance, the difference in the
temperature peaks ranged between -0.9% to 1.8% and -1.3% to +1.6% for the 1% and 2%
SRA mixtures, respectively. This indicates that the retardation of hydration process
induced by SRA is the dominant factor.
Chapter 9 294
20
25
30
35
40
45
0 6 12 18 24 30 36 42 48Age (hours)
Tem
pera
ture
(°C
)
CR1R2W4W12
(a)
20
25
30
35
40
45
0 6 12 18 24 30 36 42 48Age (hours)
Tem
pera
ture
(°C
)
R1R1W4R1W12R2R2W4R2W12
(b)
Figure 9-4: Heat of hydration for UHPC mixtures incorporating SRA and/or
wollastonite microfibers a) individually and b) combined.
Chapter 9 295
9.4.3. Degree of Hydration
Shrinkage mitigation techniques usually affect the rate of hydration reactions.
Accordingly, the degree of hydration and amount of BW will be changed (Bentz, 2000).
TGA was used to investigate the hydration evolution. The measured degree of hydration
(represented by the relative BW) versus age is shown in Fig. 9-5.
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 4 8 12 16 20 24Age (hours)
BW
(g/g
bin
der)
C W4W12 R1R1W4 R1W12R2 R2W4R2W12
Figure 9-5: Change in degree of hydration during the first day for different UHPC
mixtures.
It can be observed that all mixtures incorporating SRA exhibited lower degree of
hydration with respect to that of the control mixtures without SRA. The higher the added
SRA dosage, the lower was the degree of hydration. For instance; R1W4 and R2W4
mixtures exhibited about 18% and 27% lower amount of BW, respectively compared to
that of the W4 mixture without SRA. This is consistent with the retardation induced by
Chapter 9 296
SRA and early-age compressive strength results (see Fig. 9-1). Moreover, mixtures with
the same dosage of SRA exhibited a significant change in BW as the wollastonite
microfibers content increased from 0% to 12%. For instance, at 2% SRA, the variation in
the amount of BW between mixtures with and without wollastonite microfibers ranged
between -4.4% and +1.8%. These results confirm heat of hydration results (see Fig. 9-
4(b)).
9.4.4. Free Shrinkage and Mass Loss
The mass loss, autogenous and total shrinkage curves for mixtures incorporating SRA
and/or wollastonite microfibers are shown in Figs. 9-6 and 9-7. The measured total
shrinkage includes both drying and autogenous shrinkage (thermal deformation can be
ignored due to the small cross-section of the specimens (Baroghel-Bouny et al., 2006).
Generally, for very low w/c (Tazawa and Miyazawa, 1999), the contribution of
autogenous shrinkage to the total shrinkage can be comparable to that of drying
shrinkage.
Despite the higher porosity induced by wollastonite microfibers (Bentz, 2000), no
significant change in mass loss was observed compared to that of the control mixture
with no microfibers. This can be attributed to the discontinuity of the pore structure
induced by wollastonite microfibers (Mathur et al., 2007) and the development of greater
hydration products. This was also confirmed by MIP test results. For instance, mixture
W12 had a total porosity about 12.5% higher than that of the control mixture at 24 hrs.
However, analysis of the MIP pore size distribution data at 24 hrs, according to the
International Union of Pure and Applied Chemistry system (IUPAC) classification
(Aligizaki, 2006), showed that W12 had about 73% higher small mesopores (size range:
Chapter 9 297
2.5 to 10 nm) than that of the control mixture, while it had 6% and 8% lower large
mesopores (size range: 10-50 nm) and macropores (size > 50 nm) with respect to values
for the control mixture, respectively. The higher small mesopores content indicates more
advance development of hydration (Antonio et al., 2010). On the other hand, electrostatic
interaction between the pore wall and pore liquid can hinder the pore liquid transfer
process in pore sizes below 50 nm, while it has no effect in macrospores (Aligizaki,
2006). Hence, the lower content of macropores in W12, probably affect the water
transport mechanism, thus leading to the lower mass loss compared to that in the control
mixture.
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.00 20 40 60 80 100 120 140 160 18
Age (Hours)
Mas
s lo
ss (%
)
0
CW4W12R1R1W4R1W12R2R2W4R2W12
Figure 9-6: Mass loss for UHPC mixtures with and without different dosages of
SRA and/or wollastonite microfibers.
On the other hand, incorporating SRA slightly reduced mass loss with respect to
that of the control mixture. The higher the dosage of SRA, the greater was the reduction
Chapter 9 298
in the mass loss. Moreover, adding SRA to mixtures incorporating wollastonite
microfibers reduced the mass loss with respect to that of mixtures incorporating
wollastonite microfibers alone. For instance, mixtures R1W12 and R2W12 exhibited
about 7% and 16% lower mass loss compared to that of mixture W12 (Fig. 9-6). This can
be ascribed to the development of a drying front on the exposed top surfaces of
specimens, which in turn restricted extracting more water from deeper within specimens
(Bentz, 2006).
The general trend of total shrinkage results indicates a progressive reduction in total
shrinkage with increasing content of wollastonite microfibers, as shown in Fig. 9-7(a). A
detailed discussion on the effect of wollastonite microfibers addition on shrinkage
behaviour was given in chapter 8.
SRA induced a significant reduction in total shrinkage compared to that of the
control mixture, in agreement with previous findings (He et al., 2006, Bentz, 2006,
Powers and Brownyard, 1948). For instance, mixtures R1 and R2 exhibited about 39%
and 48% lower total shrinkage, respectively, compared to that of the control mixture at 7-
days. This can be attributed to a reduction in surface tension of the pore solution induced
by SRA (Kovler and Bentur, 2006). As a result, lower capillary stresses developed and
consequently lower total shrinkage was achieved. This reduction in capillary stresses also
explains the lower autogenous shrinkage of mixtures R1 and R2 compared to that of the
control mixture (about 41% and 59% reduction at 7-days, respectively) as shown in Fig.
9-7(b). Moreover, SRA is believed to mitigate the drop in internal relative humidity,
leading to lower self-desiccation and autogenous shrinkage (Bentz et al., 2001). A few
hours after the final setting, R1 and R2 specimens exhibited an expansion (Fig. 9-7(a)).
Chapter 9 299
This expansion played a significant role in reducing the net shrinkage along with the
reduction in autogenous shrinkage induced by SRA compared to that of control
specimens (Weiss et al., 2008). Previous study (Sant, 2009) attributed this expansion to
crystallization stresses as a result of amplifying portlandite super-saturation in the pore
solution induced by SRA addition.
Figure 9-7 shows the effect of adding SRA to mixture incorporating wollastonite
microfibers on the development of total and autogenous shrinkage. Adding SRA to
mixtures incorporating wollastonite microfibers reduced the total and autogenous
shrinkage significantly compared to that of mixtures incorporating wollastonite
microfibers alone. For instance, mixtures R1W12 exhibited about 37% and 20%
reduction in autogenous and total shrinkage, respectively compared to that of mixture
W12.
-700
-600
-500
-400
-300
-200
-100
00 20 40 60 80 100 120 140 160 180
Age (hours)
Tota
l shr
inka
ge s
trai
n (µ
m/m
)
CW4W12R1R1W4R1W12R2R2W4R2W12
(a)
Figure 9-7: Shrinkage strains for UHPC mixtures with and without different dosages of SRA and/or wollastonite microfibers under a) drying and b) sealed
conditions.
Chapter 9 300
-700
-600
-500
-400
-300
-200
-100
0
1000 20 40 60 80 100 120 140 160 180
Age (Hours)
Aut
ogen
ous
stra
in (µ
m/m
)
CW4W12R1R1W4R1W12R2R2W4R2W12
(b)
Figure 9-7 Contd’: Shrinkage strains for UHPC mixtures with and without different
dosages of SRA and/or wollastonite microfibers under a) drying and b) sealed
conditions.
The reduction in total and autogenous shrinkage compared to that of control
mixture was comparable for all mixtures incorporating the same SRA dosage, regardless
of the added amount of wollastonite microfibers. Mixtures incorporating 2% SRA had
59%, 54% and 58% reduction in autogenous strain for 0%, 4% and 12% wollastonite
microfibers content, respectively. This implies that SRA had a major effect on reducing
shrinkage, while wollastonite microfibers have a minor or negligible effect. This is due in
part to the retardation effect of SRA and early expansion. The major part of shrinkage
occurred during the first 24 hrs (Pease et al., 2005). The restraining effect of wollastonite
microfibers during this period depends on the ability to transfer shear stresses between
microfibers and the cementitious matrix, which is a function of the interfacial
Chapter 9 301
microfiber/matrix bond strength (Hameed et al., 2009, Banthia and Sheng, 1996). Hence,
the retardation of hydration reactions induced by SRA can lead to lower
microfiber/matrix bond strength and consequently lower shrinkage restraining efficiency
(Chan and Li, 1997). The pullout of wollastonite microfibers for the cementitious matrix
is shown in Fig. 9-8. On the other hand, the high early expansion induced by SRA (i.e.
between 12 and 24 hrs) could delay the development of contracting forces, which is
considered as driving force for shrinkage development (Nawa and Horita, 2004).
Figure 9-8: SEM for wollastonite microfibers pullout from cementitious matrix
incorporating SRA.
9.4.5. Restrained Shrinkage
Figure 9-9 shows the steel ring measured strains for the C, R1, R2, W4 and W12
mixtures. The SRA mixtures had a time lag in the development of strain during the 24 hrs
ring test and consequently a delay in the age at which the first crack occurred. The delay
in cracking age was about 38% and 82% for mixtures R1 and R2, respectively, compared
to that of the control mixture (Fig. 9-9(a)), which is in agreement with previous studies
(Weiss, 1997, Shah et al., 1997). This is a consequence of the slower development of the
Chapter 9 302
shrinkage for these mixtures (see Fig. 9-7). The lower shrinkage development can be
attributed to the retardation in hydration reactions and reduction in capillary stresses,
along with the early expansion which compensated for shrinkage strains and cause the
ring specimen to come out of contact with the inner steel ring thus generating no strains
(Pease et al., 2005).
Wollastonite microfibers enhanced the shrinkage cracking resistance as it delayed
the cracking age compared to that of the control mixture (Fig. 9-9(a)). A detailed
discussion on the effect of wollastonite microfibers addition on restrained shrinkage was
given in chapter 8.
Figure 9-9(b) illustrates steel ring strain results for mixtures incorporating both
SRA and wollastonite microfibers. Similar to mixtures with SRA, mixtures incorporating
wollastonite microfibers had a time lag in strain development during the ring test. This is
in agreement with previous shrinkage (see Fig. 9-7). Combining SRA and wollastonite
microfibers improved the resistance to shrinkage cracking and significantly delayed the
cracking age compared to that of the control mixture. For instance, combing 1% of SRA
and 4% of wollastonite microfibers delayed the cracking age by about 84%, 33% and
43% compared to that of the control, R1 and W4 mixtures, respectively. This represents
the net result of combining the reduction in capillary stresses induced by SRA and
improvement of cracking resistance provided by wollastonite microfibers.
Mixtures R1W4 and R1W12 exhibited comparable strain behaviour to that of
mixtures R2 and R2W4, respectively, as shown in Fig. 9-9(b). Hence, it appears that the
Chapter 9 303
SRA dosage can be reduced while achieving better or comparable cracking behaviour
with the aid of wollastonite microfibers.
Steel rings containing specimens with wollastonite microfibers tented to
experience the development of two cracks, while those with no microfibers exhibited
only one large crack. The presence of multiple cracks is an expected result, since
microfibers create internally several small restraints inside the matrix leading to lower
crack width as the cracking opening is shared by all cracks (Passuello et al., 2009). The
absence of multiple cracks in the control and SRA mixtures can be attributed to the stress
relief induced by the opening of a single crack, since, unlike matrices with fiber, there is
no ability to transfer or redistribute stresses (Shah and Weiss, 2006).
-100
-85
-70
-55
-40
-25
-10
50 24 48 72 96 120 144 168 192 216 240
Age (hours)
Ave
rage
str
ain
(µ m
/m)
CR1R2W4W12
(a)
Figure 9-9: Strain measurements of the steel ring for UHPC mixtures incorporating
SRA and/or wollastonite microfibers a) individually and b) combined.
Chapter 9 304
-100
-85
-70
-55
-40
-25
-10
50 24 48 72 96 120 144 168 192 216 240
Age (hours)
Ave
rage
str
ain
(µ m
/m)
R1R2R1W12R2W4R1W4R2W12
(b)
Figure 9-9 Contd’: Strain measurements of the steel ring for UHPC mixtures
incorporating SRA and/or wollastonite microfibers a) individually and b) combined.
9.4.6. COD Results
Washing out of SRA from concrete, especially during early-ages, can significantly
diminish its effectiveness in reducing shrinkage strains. SRA reduces the pore fluid
surface tension in concrete but does not chemically combine with other hydration
products (Rodden and Lange, 2004). Therefore, as SRA is washed out, higher shrinkage
values develop. Figure 9-10(a) shows shrinkage results for the control, R2, R2W4 and
R2W12 mixtures under submerged conditions.
Chapter 9 305
-100
0
100
200
300
0 24 48 72 96 120 144 168
Age (hours)
Stra
in (µ
ε)
CR2R2W4R2W12
(a)
-200
-150
-100
-50
00 24 48 72 96 120 144 16
Age (hours)
Stra
in (µ
ε)
8
CR2R2W4R2W12
(b)
Figure 9-10: Measured strains under submerged condition for control, SRA and/or
wollastonite microfibers UHPC mixtures: a) observed and b) initiated to zero after
the early expansion.
Chapter 9 306
Under submerged conditions, specimens continued to swell during the first 24 hrs,
then started to shrink until the end of the investigated period. The early swelling resulted
in a lower net shrinkage compared with that of sealed specimens (see Fig. 9-10(a)). The
swelling of submerged specimens can be ascribed to the continuous supply of water,
allowing the calcium silicate hydrate (CSH) to absorb water and expand (Neville, 1996).
The reduction in the swelling strain with time can be explained by the fact that the
progress of the hydration reactions reduced and depercolated the concrete capillary
porosity, which interfered with water imbibing from the surrounding environment.
Hence, no further expansion took place, while internal self-desiccation occurred as water
was consumed by the hydration process. Mixtures incorporating wollastonite microfibers
exhibited lower early-age expansion compared to that of the control and SRA mixtures.
Moreover, the higher the wollastonite microfibers content, the lower was the early
expansion. This can be attributed to the discontinuity of the pore structure induced by
wollastonite microfibers (Mathur et al., 2007) which probably affected water movement
within the concrete (Passuello et al., 2009).
In order to capture the shrinkage behaviour under submerged conditions, the
strain curves were initiated to zero at the end of the swelling, as shown in Fig. 9-10(b).
Mixture R2 had a comparable shrinkage strain to that of the submerged control
specimens with no SRA, while mixtures R2W4 and R2W12 exhibited a lower shrinkage
strain. Moreover, the reduction in shrinkage strain increased with higher content of
wollastonite microfibers. It is hypothesized that SRA is washed out with migrating water
from SRA mixtures, jeopardizing their shrinkage mitigation efficiency, while lower
amounts of SRA were washed out from R2W4 and R2W12 mixtures.
Chapter 9 307
To examine this hypothesis, chemical oxygen demand (COD) tests were
conducted on water samples taken from the submersion tanks. The COD results in Fig. -
11 for the C, R2, R2W4 and R2W12 mixtures confirm the existence of SRA in the
submerging water. The cumulative level of leached SRA increased with time. However,
mixtures R2W4 and R2W12 had lower COD values compared to that of mixture R2,
which indicates a reduction in the amount of SRA washed out. The higher the content of
wollastonite microfibers in specimens, the lower was the measured COD. This can be
ascribed to the discontinuity of the pore structure of the wollastonite microfibers, which
interfered with the washing out of SRA. Hence, a higher SRA content remained inside
the pore solution, thus reducing capillary stresses induced by the internal self-desiccation
and leading to lower shrinkage.
0
500
1000
1500
2000
2500
1 3 5 7Age (days)
CO
D v
alue
(mg/
L)
0
5
10
15
20
25
Cum
ulat
ive
SRA
loss
(%)
ControlSRAR2W4R2W12SRA loss
R2
Figure 9-11: COD values for control, SRA and/or wollastonite microfibers UHPC
mixtures.
Chapter 9 308
9.6. CONCLUSIONS
In this study, the shrinkage of UHPC mixtures with and without SRA and/or wollastonite
microfibers was investigated under sealed, drying and submerged conditions. Based on
this work, the following conclusions can be drawn:
1) Wollastonite microfibers addition in UHPC mixtures had a positive effect on the
early-age compressive strength.
2) The addition of wollastonite microfibers enhanced compressive strength of UHPC
mixtures incorporating SRA.
3) Incorporating both wollastonite microfibers and SRA in UHPC delayed the time
of cracking through retarding the development of shrinkage strains and resisting
the coalescence of micro-cracks.
4) The washout of SRA under submerged conditions can jeopardize its effectiveness
in mitigating shrinkage strains.
5) Incorporating wollastonite microfibers in UHPC appears to promote pore
discontinuity, thus leading to lower mass loss, less drying shrinkage and reduced
SRA washing out.
Chapter 9 309
9.7. REFERENCES
Aligizaki, K.K., (2006). Pore Structure of Cement-based Materials: testing, interpretation and requirements. 1St Edition, Taylor & Francis, New York, USA, 388 p.
Antonio, A., Melo, N., Maria, A.C. and Wellington, R., (2010), “Mechanical properties,
drying and autogenous shrinkage of blast furnace slag activated with hydrated lime and gypsum,” Cement and Concrete Composites, Vol. 32, No. 4, pp. 312-318.
Banthia, N. and Sheng, J., (1996), “Fracture toughness of microfiber reinforced cement
composites,” Cement and Concrete Composites, Vol. 18, No. 4, pp. 251-269. Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafai, N., (2006),
“Autogenous deformations of cement pastes: Part II. w/c effects, micro-macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp.123-136.
Bentz, D.P., (2000), “Fibers, percolation, and spalling of high performance concrete,”
ACI Materials Journal, Vol. 97, No. 3, pp. 351-359. Bentz, D.P., (2006), “Influence of Shrinkage-Reducing Admixtures on Early-Age
Properties of Cement Pastes," Journal of Advanced Concrete Technology, Vol. 4, No. 3, pp. 423-429.
Bentz, D.P., Geiker, M.R. and Hansen, K.K., (2001), “Shrinkage-reducing admixtures
and early-age desiccation in cement pastes and mortars,” Cement and Concrete Research, Vol. 31, No. 7, pp. 1075-1085.
Chan, Y.W. and Li, V.C., (1997), “Age effect on the characteristics of fiber/cement
interfacial properties,” Journal of Materials Science, Vol. 32, No. 19, pp. 5287-5292.
Eberhardt, A.B. and Kaufmann, J., (2006), “Development of shrinkage reduced self-
compacting concrete,” ACI Journal, Special addition NO. P-235, pp. 13-30. Hameed, R., Turatsinze, A., Duprat, F. and Sellier, A., (2009), “Metallic fiber reinforced
concrete: effect of fiber aspect ratio on the flexural properties,” ARPN Journal of Engineering and Applied Sciences, Vol. 4, No. 5, pp. 67-72.
He, Z., Li, Z.J., Chen, M.Z. and Liang, W.Q., (2006), “Properties of shrinkage-reducing
admixture-modified pastes and mortar,” Materials and Structures, Vol. 39, No. 4, pp. 445-453.
Kovler, K. and Bentur, A., (2009), “Cracking sensitivity of normal and high strength
concretes,” ACI Materials Journal, Vol. 106, No. 6, pp. 537-542.
Chapter 9 310
Lea, F.M., (1988). Lea's Chemistry of Cement and Concrete. 4th Edition, (ed.) P.C. Hewlett, J. Wiley, New York, 1053 p.
Low, N.M.P. and Beaudoin, J.J., (1992), “Mechanical properties of high performance
cement binders reinforced with wollastonite microfibers,” Cement and Concrete Research, Vol. 22, No. 5, pp. 981-989.
Mathur, R., Misra, A.K. and Goel, P., (2007), “Influence of wollastonite on mechanical
properties of concrete,” Journal of Scientific and Industrial Research, Vol. 66, No. 12, pp.1029-1034.
Nawa, T. and Horita, T., (2004), “Autogenous shrinkage of High- performance concrete,”
Proceedings of the International Workshop on Microstructure and Durability to Predict Service Life of Concrete Structures, Sapporo, Japan.
Neville, A.M., (1996). Properties of Concrete, 4th Edition, John Wiley &Sons, New York, 797 p.
Nmai, C., Tomita, R., Hondo, F. and Buffenbarger, J., (1998), “Shrinkage reducing admixtures,” Concrete International, Vol. 20, No. 4, pp. 31-37.
Passuello, A., Moriconi, G. and Shah, S.P., (2009), “Cracking
behavior of concrete with shrinkage reducing admixtures and PVA fibers,” Cement and Concrete Composites, Vol. 31, No. 10, pp. 699-704.
Pease, B.J., Shah, H.R., and Weiss, W.J., (2005), “Shrinkage behavior and residual stress
development in mortar containing shrinkage reducing admixture (SRA’s),” ACI-Special Publication on Concrete Admixtures, Vol. 227, pp. 285–302
Powers, T.C. and Brownyard, T.L., (1948), “Studies of the physical properties of
hardened portland cement paste,” Bulletin No. 22, Research Laboratories of the Portland Cement Association. Reprinted from Journal of the American Concrete Institute, October 1946–April 1947, Proceedings 43, Detroit, pp. 971-992.
Rajabipour, F., Sant, G. and Weiss, J., (2008), “Interactions between shrinkage reducing
admixtures (SRA) and cement paste's pore solution,” Cement and Concrete Research, Vol. 38, No. 5, pp. 606-615.
Rodden, R.A. and Lange, D.A., (2004) “Feasibility of shrinkage reducing admixtures for
concrete runway pavements,” Technical Note No. 4, The Center of Excellence for Airport Technology (CEAT), 6 p.
Sant, G.N., (2009), “Fundamental investigations related to the mitigation of volume
changes in cement-based materials at early ages,” Ph.D. thesis, Purdue University, 227 p.
Chapter 9 311
Schmidt, M., (1992), “Cement with inter-ground additives: capabilities and environmental relief (I),” Zement Kalk Gips, Vol. 45, No. 7, pp. 64-69.
Shah, S. P., Weiss, W. J., and Yang, W., (1997), “Shrinkage cracking in high
performance Concrete." Proceedings of the PCI/FHWA International Symposium on High Performance Concrete, New Orleans, Louisiana, pp. 217-228.
Shah, S.P., Weiss, W.J. and Yang, W., (1998), “Shrinkage cracking-can it be
prevented?,” Concrete International, Vol. 20, No. 4, pp. 51-55. Shah, H. R. and Weiss, J., (2006), “Quantifying shrinkage cracking in fiber reinforced
concrete using the ring test,” Materials and Structures, Vol. 39, No.9, pp. 887-899.
Tazawa, E. and Miyazawa, S., (1999), “Effect of constituents and curing condition on
autogenous shrinkage of concrete,” Proceedings of the International Workshop Autogenous Shrinkage of Concrete, (ed.) Eichi Tazawa, Taylor & Francis, Hiroshima, Japan, pp. 269-280.
Weiss, J., Lura, P., Rajabipour, F. and Sant, G., (2008), “Performance of shrinkage-
reducing admixtures at different humidities and at early ages,” ACI Materials Journal, Vol. 105, No. 5, pp. 478-486.
Weiss, J., (1997), “Shrinkage Cracking in restrained Concrete Slabs: Testing methods,
material compositions, shrinkage reducing admixture and theoretical modeling,” M.S. thesis, Northwestern University, Evanston, Illinois, USA, 128 p.
Chapter 10 312
CHAPTER TEN
ARTIFICIAL NEURAL NETWORK MODELING OF EARLY-AGE AUTOGENOUS SHRINKAGE OF
CONCRETE*
In this chapter, an artificial neural networks (ANN) model for the early-age autogenous
shrinkage of concrete is proposed. The model inputs include the cement content, water-
to-cement ratio (w/c), type and percentage of supplementary cementitious materials, total
aggregate volume, curing temperature, and hydration age. The autogenous shrinkage of
concrete is the model output. Subsequent to model validation using experimental results
of chapter 3, a parametric study was carried out to identify the effects of input variables
on the evolution of autogenous shrinkage. Moreover, the autogenous shrinkage database
assembled and used in the training of the proposed ANN is considered as a contribution
to the state-of-the-art knowledge; it includes testing results on modern concretes under
various environmental conditions.
10.1. INTRODUCTION
Concrete is the most widely used construction material world-wide (Vanikar, 2004).
Shrinkage of concrete initiates cracks that can lead to the ingress of aggressive
substances, which jeopardizes the durability of concrete structures (Lura et al., 2007).
Hence, concrete shrinkage has been a matter of great concern in the design of durable
concrete infrastructure (Toutanji et al., 2004). Shrinkage occurring over a long time
*A version of this chapter has been submitted for review to the ACI Materials Journal.
Chapter 10 313
period is attributed to drying. Early-age drying shrinkage induces internal tensile stresses
while the concrete has not yet gained sufficient tensile strength to withstand such stresses,
which leads to cracking (Sivakumar and Santhanam, 2006).
At early-ages, another type of concrete shrinkage (so called “autogenous
shrinkage”) occurs without moisture exchange with the environment (Bentz and Jensen,
2004). Autogenous shrinkage is a mechanism whose importance has grown with the
increasing use of low w/c concrete. For many years, autogenous shrinkage, which can
mainly be attributed to internal self-desiccation due to the hydration of cement (Bentz
and Jensen, 2004), had been neglected as it caused minor strains in conventional concrete
members compared to that caused by drying shrinkage (Zhang et al., 2003). However,
autogenous shrinkage was found to be too large to be neglected and to have critical
effects on low w/c concrete members (Igarashi et al., 2000). The evolution of the
autogenous shrinkage of concrete depends on many factors, including the w/c (Zhang et
al., 2003), aggregate content, cement content (Tazawa et al., 1995), curing temperate
(Mak et al., 1999), and type and percentage of supplementary cementitious materials
(Zhang et al., 2003). Such parameters are highly interdependent and exhibit complex
combined roles on the development of autogenous shrinkage.
Existing models for drying shrinkage do not normally perform well for
autogenous shrinkage of modern concretes (ACI 209R-92 (ACI Committee 209, 1992),
Model B3 (Bažant, 1995), CEB-FIP 1990 (CEB-FIP 1990, 1991), Gl2000 (Gardner and
Lockman, 2001)), since they are limited to concretes with mean 28-days compressive
strengths ranging from 20 to 70 MPa. They also cannot capture the effects of high
Chapter 10 314
contents (more than 30%) of silica fume, fly ash and other natural pozzolans, and they do
not take into account the faster evolution of properties at early-ages in very low w/c ratio
systems (ACI Committee 209, 2008, Lpَez and Pacios, 2006). Such models generally use
mathematical formulae relating variables by means of statistical coefficients calculated
from certain databases. The characteristics of the databases, including component
materials, age and environmental conditions, limit the validity and application range of
such models (Lpَez and Pacios, 2006).
On the other hand, ANN is a powerful modeling tool for problems where the rules
which govern the results are either not defined properly or too complex (Adeli, 2001,
Flood and Kartam, 1994). A number of applications of ANN in concrete materials and
structures have been proposed by several researchers (Bilgehan and Turgut, 2010,
Rattapoohm and Pichai, 2009, Yeh, 2008, El-Chabib et al., 2005). ANN is an artificial
intelligence technique that does not require mathematical relationships between variables.
It has learning, self-organizing and auto-improving capabilities (Demir, 2008) allowing it
to capture complex interactions among variables without any previous knowledge of the
nature of these interactions. A properly trained ANN also has the ability to recall full
patterns from incomplete or noisy data (Rafiq et al., 2001). Due to such exceptional
capabilities, ANN has been used in a wide range of engineering applications.
This chapter demonstrates the potential for using artificial neural networks to
predict the autogenous shrinkage of concrete under different curing temperatures. The
assembled database, model architecture and training process of the ANN network are
described. Moreover, the influence of important parameters including w/c,
superplasticizer dosage and curing temperature on autogenous shrinkage were
Chapter 10 315
quantitatively analyzed and evaluated with respect to knowledge available in the open
literature.
10.2. RESEARCH SIGNIFICANCE
For many years, the autogenous shrinkage of concrete associated with the internal
consumption of water during cement hydration and the associated self-desiccation had
not been regarded as a serious problem. However, with the increasing use of low w/c
concrete, the importance of autogenous shrinkage has grown since it is too large to be
neglected. Autogenous shrinkage can have a comparable value to that of drying
shrinkage. Therefore, predicting the autogenous shrinkage is an important aspect in the
service life analysis and design of durable concrete structures. In this study, an attempt
was made to develop an artificial intelligence-based tool for accurately predicting the
autogenous shrinkage of concrete based on its mixture design. A powerful modeling tool,
artificial neural networks, was employed on a comprehensive database assembled from
the open literature. The ANN model allows predicting autogenous shrinkage along with
the influence of key parameters on the autogenous shrinkage behaviour of concrete.
10.3. NEURAL NETWORK APPROACH
ANN has been used to relate the autogenous shrinkage of concrete and its mixture
composition at different curing temperatures. ANN learns from input information similar
to the operation of a biological brain (Haykin, 1999). It has the capability of
generalization, classification, pattern recognition, function approximation and simulation
of sophisticated operations (Haykin, 1999). ANN structure consists of parallel multiple
layers of linear and nonlinear processing elements (i.e. neurons) which can be classified
Chapter 10 316
into: an input layer, an output layer, and hidden layers (Demir, 2008), as shown in Fig.
10-1. These neurons are linked by variable weights. The input layer receives original data
(xi), which is adjusted by connection weights (wij) and biases (b). The adjusted inputs are
subjected to a summation process to form a single input (net) for all inputs received from
the input laye
j
r (Eq. 10-1) (Sarıdemir, 2009).
bxw)net(n
1iiijj +⋅=∑
= Eq. 10-1
This single input is modified by an activation function (f(x)) to generate an output value
of the processing unit through the hidden layers. The error between the network outputs
and desired targets is calculated and then propagated back to the network through a
learning algorithm. The implementation of such an algorithm updates the network
weights and biases in the direction in which the total network error decreases rapidly.
ANN then synthesizes and memorizes the relationship between the inputs and outputs
through a training process. Hence, the used data in the training process should be
sufficient and representative to allow the ANN to recognize the underlying structure of
the information involved. Once an ANN is established and well-trained, it will be capable
of predicting outputs of any input set of data, and predicting the outcome of any
unfamiliar set of input located within the range of the training data with an acceptable
degree of accuracy.
In civil engineering, feed-forward neural networks along with back-propagation
algorithms are widely used and have shown efficient performance (Demir, 2008,
Sarıdemir, 2009, Venkiteela et al., 2010, Adhikary and Mutsuyoshi, 2006, Mukherjee and
Chapter 10 317
Biswas, 1997). Hence, they were selected for constructing the proposed ANN model in
this study.
Figure 10-1: The architecture of ANN models.
10.3.1. Feed-Forward Neural Network
Feed-forward neural network (FFN) is the most widely used model in engineering
applications. In FFN, neurons are arranged in layers and all the neurons in each layer are
linked to all the neurons in the next layer (Adhikary and Mutsuyoshi, 2006). In general,
the FFN consists of one input layer, one output layer and one or more hidden layers of
neurons (Mukherjee and Biswas, 1997). The phrase “Feed-forward” indicates that the
data move forward from one layer to the next during ANN modeling (Venkiteela et al.,
2010). The input layer receives input information and passes it forward to the neurons of
the hidden layer, which in turn passes the information to the output layer. The output
from the output layer is the corresponding prediction of the model for the data set
Chapter 10 318
supplied at the input layer. To construct a stable FFN for a particular problem, the
optimum number of neural units in each layer was selected using a trial and error
approach as recommended by (Rafiq et al., 2001).
10.3.2. Back-Propagation Learning Algorithm
Learning algorithms are techniques used to establish connections (i.e. weights and biases)
between neurons forming the network structure and to adjust both weights and biases to
obtain the desired values. There are two broad categories of algorithms: unsupervised
(weights and biases are modified in response to network inputs only) and supervised
(weights and biases are modified in order to move the network outputs closer to the
targets) (Haykin, 1999).
One of the well-known supervised training algorithms for the FFN is the back-
propagation algorithm. In this algorithm a gradient descent technique is applied to
minimize the error for a particular training pattern in which it adjusts the weights by a
small amount at a time (Sarıdemir, 2009, Venkiteela et al., 2010, Adhikary and
Mutsuyoshi, 2006, Mukherjee and Biswas, 1997, Topçu and Sarıdemir, 2007, Ince,
2004). The learning error is calculated using the following equation (Eq. 10-2)
∑ −=i
2ii2
1 )o(tError Eq. 10-2
Where ti is the target output and oi is the predicted output at neuron (i), respectively. In
the back-propagation phase, the error between the predicted and target output values is
calculated and used to update the weights between neurons using the following equation
(Eq. 10-3) (Ince, 2004):
Chapter 10 319
)1t(wo)t(w j,iijj,i −∆β+ηδ=∆ Eq. 10-3
where is the weight change at the end of iteration (t), local error gradient,
is the weight change at iteration (t-1),
)t(w j,i∆ jδ
)1t(w −∆ j,i η , β are learning and moment rate,
respectively. Hence, the error serves as a feedback for the learning process. The iteration
continues until a satisfactory convergence is achieved (Mukherjee and Biswas, 1997).
10.4. PROPOSED ANN MODEL
A multilayered feed-forward neural network with a back-propagation algorithm was used
to model the autogenous shrinkage behaviour in the present study. Two hidden layers
with a Tansigmoid (tansig Eq. 10-4) transfer functions were used, while a pure linear
transfer function (Eq. 10-5) was used in the output layer.
1e12nsig
n2−
+=
− )()(tan Eq. 10-4
n(n)purelin = Eq. 10-5
During the training process, the design of network architecture starts with fewer hidden
neurons, and then the number of hidden neurons is adjusted by assessing the network
error through applying a trial and error approach (Nehdi et al., 2007). In order to improve
generalization, the regularization method was applied. This involves modifying the
performance function, which is normally chosen to be the sum of squares of the network
errors on the training set (Demuth and Beal, 1998). The new performance function
measures the network performance as the weight sum of two factors: the mean squared
error and the mean squared weight and bias values (Eq. 10-6). Using this modified
Chapter 10 320
performance function causes the network to have smaller weights and biases, which
forces the network response to be smoother and less likely to over fit.
γ)msw(1γmsemsereg −+= Eq. 10-6
Where (msereg) is the mean squared error with a regularization performance function,
( γ ) is the performance ratio which gives weight to the mean square errors (mse) (Eq. 10-
7) and the mean square weights (msw) (Eq. 10-8).
∑=
−=N
1i
2ii )o(t
N1mse Eq. 10-7
Where N is the number of iterations, ti is the target output and oi is the predicted output at
neuron (i), respectively.
∑=
=n
1j
2jw
n1msw Eq. 10-8
Where n is the number of neurons, w is the weight value for neuron (j).
To simplify the learning process and reduce the required time for training, the
back-propagation Levenberg-Marquardt Algorithm (LMA) was adopted as the learning
algorithm (Nehdi et al., 2001). The LMA operates in a batch mode at which the weights
and biases of the network are updated only after the entire training set has been applied to
the network (Demuth and Beal, 1998). LMA propagates back the errors computed at the
output layer to the network based on the Jacobian matrix J, which contains the first
derivatives of the network errors with respect to weights and biases. An iteration of such
algorithm can be written as follows (Eq. 10-9)
Chapter 10 321
[ ] eJµIJJww T1Tj1j
−+ +−= Eq. 10-9
where wj is a vector of current weights and biases, µ is a learning rate, J is the Jacobian
matrix, is the transpose matrix of J, I is the identity matrix, and e is a vector of
network errors. Parameters of the established ANN model are shown in Table 10-1.
TJ
The available set of data is divided into three subsets, namely, training, validation,
and testing. The training data is used to train the model to recognize the patterns between
input and output data. The validation data is used to evaluate the effectiveness of the
designed model in generalizing the underlying relationships and achieving a good
performance when new data are introduced. The final model is tested with the testing
data set, not presented to the model before, to ensure that predictions are real and not
artifacts of the training process (Demuth and Beal, 1998). Before training, all data (i.e.
inputs and targets) were scaled so that they fall in the range [-1,1]. This pre-processing
step increases the efficiency of the ANN training (Rafiq et al., 2001).
Table 10-1: The values of parameters used in the ANN model.
Parameters ANN Number of input layer neurons 8 Number of hidden layer 2 Number of first hidden layer neurons 12 Number of second hidden layer neurons 8 Number of output layer neurons 1 Minimum gradient 1×10‐10
Goal 1×10‐5
Chapter 10 322
10.5. MODEL DATABASE
The ability of the ANN model to predict the autogenous shrinkage behaviour of concrete
mixtures will largely depend on how comprehensive the database is. In other words, it
will depend on the availability of experimental data that are capable to teach the ANN
relationships between the concrete mixture variables and its measured autogenous
shrinkage. An extensive literature review identified a great amount of published data on
concrete shrinkage. However, in many cases there is insufficient information regarding
the exact concrete mixture composition, testing methodology, curing conditions,
incomplete or graphically presented data, and the absence of other important parameters
needed to properly characterize autogenous shrinkage (e.g. age at which the measuring of
autogenous shrinkage started). The exclusion of such incomplete data reduced the extent
of the database available for training the network.
It should also be highlighted that autogenous shrinkage of concrete has not been
researched as extensively as drying shrinkage until Tazawa and Miyazawa (Tazawa and
Miyazawa, 1993) emphasized its importance for low w/c mixtures. To avoid further
complexity, only experimental data having mixture components with comparable
physical and chemical properties were identified for the training and testing of the
network. Using the aforementioned criteria, a number of data sets were selected from
different studies and used to train, validate and test the ANN model, as summarized in
Table 10-2. For ANN training, eight input variables were selected: w/c, cement content,
silica fume percentage, fly ash percentage, superplasticizer content, total aggregate
volume, curing temperature and hydration age. The measured autogenous shrinkage of
concrete was the single output.
Chapter 10 323
Table 10-2: Database sources and variables range of input and output.
Source No. of
mixtures
Variables
Maxim
um
Minim
um
Unit
1 Zhang et al., 2003 9 2 Igarashi et al., 2000 4
Cement 924 292 kg/m3
3 Lura et al., 2007 1 4 Mak et al., 1999 4
w/c 0.60 0.16 ‐‐‐‐‐
5 Yang et al., 2005 3 6 Loukili et al.,2000 1
SF 30 0.0 % cement
7 Bentur et al., 2001 1 8 Lee et al., 2006 4
FA 50 0.0 % cement
9 Akkaya et al., 2007 3 10 Pipat et al., 2005 5
TA 72 40 % total volume
11 Yun et al., 2006 4 12 Mazloom et al., 2004 4
SP 4 0 % cement
13 Meddah and Sato, 2010 1 14 Nassif et al., 2007 5
Temp. 65 10 °C
15 Konsta et al., 2003 3 16 Meddah et al., 2006 7
Age 360 0.5 hours
17 Guangcheng et al., 2006 4 18 Akcay and Tasdemir,2009 1
Autogenous shrinkage
‐1012 +20 με
19 Daniel and Ted, 2008 1 20 Ma et al., 2003 4 21 Brook et al., 1999 8 Total No. of mixtures 77
TA: Total aggregate volume; SF: Silica fume; FA: Fly ash; SP: Superplasticizer
10.6. RESULTS AND DISCUSSION
Since there is no precise method for partitioning the database, the ANN model was
trained with randomly selected 60% of the total database, while 20% was used for
validation and the other 20% for testing (West et al., 1997). The performance of an ANN
model depends on the success of the training process. A successfully trained ANN model
should give accurate output predictions, not only for input data used in the training
Chapter 10 324
process, but also for new testing data unfamiliar to the model within the range of the
training database. Moreover, very good ANN models normally have only slight
difference between their validating and testing errors (Amegashie et al., 2006). At the
training stage, the performance of the ANN model was assessed statistically based on
root-mean-squared (RMS) error, absolute fraction of variance (R2), and mean absolute
percentage error (MAPE) between model and experimental results which are expressed in
the following equations (Eq. 11,12 and 13) , respectively (Sarıdemir, 2009, Topçu and
Sarıdemir, 2008)
∑=
−=n
1i
2iin
1 )o(tRMS Eq. 11
( )
( ) ⎟⎟⎟⎟⎟
⎠
⎞
⎜⎜⎜⎜⎜
⎝
⎛−
−=
∑
∑
=
=n
1i
2i
n
1i
2ii
t
ot12R
Eq. 12
∑
∑
=
=
−=
n
1ii
n
1iii
n1
t
otMAPE
Eq. 13
Where ti is the target output, is the predicted output, and n is the number of data points. io
Satisfactory performance of the training process was verified by requiring the
ANN model to predict the autogenous shrinkage of concrete mixtures from the training
data set using the eight input variables. Predictions of the ANN model are shown in Fig.
10-2. The figure includes the equity line, as a reference, which represents the condition of
equal values for the predicted and measured autogenous shrinkage strains. It can be noted
Chapter 10 325
that the ANN model had captured the input-output relationships since the points are
mostly located on or slightly under/above the equity line between the experimental and
predicted expansion values. The RMS, R2 and MAPE values were 29 µε, 0.98 and 12.2%,
respectively, which indicates that the performance of ANN is satisfactory.
To examine the generalization capacity of the ANN, it was tested using the testing
data (20% of the original database). Such testing points were not previously presented to
the model, and thus the predictive capacity of the model for new data can be evaluated.
The eight input parameters of the testing data points were introduced to the ANN model
and the response (predicted autogenous shrinkage) is shown in Fig. 10-2. Similar to the
case of the training data, the model achieved good predictions relative to the actual
experimental data; testing data points were mostly located on or slightly deviated from
the equity line (Fig. 10-2). The RMS, R2 and MAPE of the testing data points were 42 µε,
0.965 and 15.1%, respectively. Hence, it can be deduced that the ANN had a satisfactory
generalization capacity for predicting the autogenous shrinkage of similar concrete
mixtures exposed to various curing temperatures. In addition, statistical parameters for
the model validation data were comparable to that of the training and testing data (i.e.
RMS =39 µε, R2=0.95 and MAPE= 13.2%) indicating an excellent performance of the
model (Shang et al., 2004).
Chapter 10 326
-1200
-1000
-800
-600
-400
-200
0
200
-1200 -1000 -800 -600 -400 -200 0 200Target autogenous shrinkage (µε)
Pred
icte
d au
toge
nous
shr
inka
ge (µ
ε) TrainingValidating TestingEquity Line
Figure 10-2: Response of ANN model in predicting autogenous shrinkage of
concrete.
10.6.1. Validating ANN Using Experimental Work
To demonstrate the utility of the proposed ANN model, experimental measurements from
Chapter 3 for UHPC mixtures with w/c=0.22 and 0.25 cured at different temperatures,
namely, 10, 20, 40 °C were collected and compared to that predicted by the trained ANN
model. Details for the experimental program, materials, specimens and testing procedure
were given in Chapter 3.
The measured and predicted autogenous shrinkage curves are illustrated in Fig.
10-3. It can be observed that the ANN predictions were in good agreement with the
measured results throughout the entire range of the shrinkage behaviour. A comparison
between the measured and ANN model-predicted autogenous shrinkage values shows a
Chapter 10 327
deviation of about ±20%, which is quite reasonable according to (ACI Committee 209,
2008) (Fig. 10-4).
-700
-600
-500
-400
-300
-200
-100
0
1000 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
Measured-10°C Predicted -10°CMeasured-20°C Predicted -20°CMeasured-40°C Predicted -40°C
(a)
-700
-600
-500
-400
-300
-200
-100
0
1000 20 40 60 80 100 120 140 160 180
Age (hours)
Aut
ogen
ous
Stra
in (µ
ε)
Measured-10°C Predicted-10°CMeasured-20°C Predicted-20°CMeasured-40°C Predicted-40°C
(b)
Figure 10-3: Measured autogenous shrinkage values compared with predicted a)
w/c=0.22 and b) w/c=0.25.
Chapter 10 328
-700
-600
-500
-400
-300
-200
-100
0-700-600-500-400-300-200-1000
Measured strain(microstrain)
Pred
icte
d st
rain
(mic
rost
rain
)
w/c=0.22-T=10°Cw/c=0.22-T=20°Cw/c=0.22-T=40°Cw/c=0.25-T=10°Cw/c=0.25-T=20°Cw/c=0.25-T=40°CEquity Line
Figure 10-4: Deviation between measured and predicted autogenous shrinkage
values.
Moreover, residual analysis was applied to evaluate the performance of the ANN
models. The residual value R is defined as the difference between the measured and
ANN-predicted shrinkage strains (Eq. 10-14).
R =Predicted value – Measured value Eq. 10-14
The residuals of tested mixtures at the different curing temperature are plotted in Fig. 10-
5. Positive residuals indicate that the model underestimated the shrinkage strains and
negative residuals indicate that the model overestimates them. The distributions of the
residuals in the negative region are greater than that in the positive range, indicating that
the ANN model slightly overestimated autogenous shrinkage. However, about 97% of
residuals are within a low range of ±100 µε, which indicates a reasonable performance of
the ANN model.
Chapter 10 329
-400
-300
-200
-100
0
100
200
300
400
0 50 100 150 200Age (hours)
Res
idua
ls (µ
ε)
w/c=0.22-T=10°C w/c=0.25-T=10°Cw/c=0.22-T=20°C w/c=0.25-T=20°Cw/c=0.22-T=40°C w/c=0.25-T=40°C
Figure 10-5: Residual values for ANN model.
10.6.2. Parametric Analyses
The ANN model showed a satisfactory performance and demonstrated its ability to
predict autogenous shrinkage strains for a wide range of concrete mixture designs under
different curing temperatures. The purpose of this section is to exploit the abilities of the
ANN model in capturing the influence of individual input parameters on autogenous
shrinkage development. The analysis was done by randomly selecting a concrete mixture
to create multiple new mixtures (not previously presented to the model) with various
levels (within the range of training data) of the parameter of interest. Among the model
inputs, the w/c, superplasticizer content and curing temperature were selected due to
limited and/or conflicting results about their effects on autogenous shrinkage.
Chapter 10 330
10.6.2.1 Effect of Water-to-Cement Ratio
The w/c has a great influence on the autogenous shrinkage behaviour of a concrete
mixture. To investigate changes in the autogenous shrinkage magnitude due to variation
of the w/c ratio, the w/c ratio was changed while the water amount was held constant.
This corresponds to an increasing cement content and paste content while the aggregate
amount remains constant. A constant dosage of superplasticizer was added; hence
workability of the tested mixtures varied. As expected, the mixture with the lowest w/c
ratio had the greatest amount of autogenous shrinkage. Moreover, autogenous shrinkage
increased with a drop of the w/c as shown in Fig. 10-6. Increasing the cement content
increases chemical shrinkage, which is a main driving force behind autogenous
shrinkage. In agreement with previous work (Baroghel-Bouny et al., 2006), the
magnitude of the predicted autogenous shrinkage at a given age increased linearly as the
w/c decreased. The w/c is likely to be below 0.25 for ultra-high-performance concrete
(UHPC). However, limited data is available on the relationship between the autogenous
shrinkage behaviour and w/c for w/c values below 0.25. Therefore, with the aid of the
ANN generalization capability, this relationship was successfully investigated.
Below w/c = 0.24, the relationship of autogenous shrinkage and w/c exhibited a
similar linear trend but with a higher slope. This indicates that the rate of autogenous
shrinkage development is faster and more sensitive to w/c ratio reduction below 0.25
compared to that at higher w/c (> 0.25). Reviewing the limited test results for mixtures
with w/c below 0.25, a good agreement between model predictions and experimental data
was found. For instance, results reported by (Tazawa et al., 2000) show that reducing the
w/c from 0.23 to 0.17 led to about 1.5 times higher rate of autogenous shrinkage
Chapter 10 331
compared to that induced by reducing the w/c from 0.4 to 0.3. Therefore, a polynomial
relationship was proposed to capture the behaviour over a wider w/c range from 0.16 to
0.44 (Fig. 10-7).
-600
-500
-400
-300
-200
-100
0
1000 12 24 36 48 60 72
Age (hours)
Aut
ogen
ous
shrin
kage
(µε)
w/c=0.18 w/c=0.22 w/c=0.26 w/c=0.30w/c=0.34 w/c=0.38 w/c=0.42
Figure 10-6: Sensitivity of ANN model to the w/c in predicting autogenous
shrinkage.
Chapter 10 332
y = -5429.6x2 + 4741.9x - 1106.8R2 = 0.9868
y = -3820.7x2 + 3520.2x - 830.48R2 = 0.9889
-600
-500
-400
-300
-200
-100
00 0.08 0.16 0.24 0.32 0.4 0.48
w/c
Aut
ogen
ous
shrin
kage
(µε)
72 hrs
24 hrs
Figure 10-7: Correlation between autogenous shrinkage of concrete measured at
T=20°C and w/c at ages 24 and 72 hours.
10.6.2.2 Effect of Superplasticizer
Superplasticizer (SP) is a key component in modern concretes. It is used to improve
workability and allow for adequate placement on site. In the literature, there is conflicting
observations about the effect of superplasticizer on autogenous shrinkage. An increasing
trend was reported (e.g. Holt and Leivo, 2004, Holt, 2005), as well as a reducing trend
(e.g. Tazawa and Miyazawa, 1995, Nawa and Horita, 2004). In order to demystify such
discrepancies, the ANN model was used to capture the effect of the SP dosage on the
autogenous shrinkage behaviour. Figure 10-8 illustrates the effect of the SP amount on
the autogenous shrinkage of concrete mixtures with a w/c = 0.26 used with various
cement contents. It can be observed that the autogenous shrinkage of concrete increased
Chapter 10 333
with increasing SP dosage until reaching a threshold dosage, beyond which autogenous
shrinkage decreased. This threshold dosage increased with increasing cement content.
This behaviour can be explained as follows: Chemical shrinkage, which is a
reduction in the volume of hydration products compared to that of the reacting
constituents, is a main driving mechanism behind autogenous shrinkage (Tazawa et al.,
1995). Adding SP deflocculated cement particles, leading to better dispersion, and
consequently faster rate of hydration reactions. This ultimately led to a higher rate of
autogenous shrinkage (Holt and Leivo, 2004, Holt, 2005). Conversely, a higher SP
dosage was found to have the side effect of slowing the chemical activity of hydration
through provoking an ionic double layer around cement grains (Morin et al., 2001).
Therefore, it can be argued that as long as the added SP dosage is less than the threshold
dosage, SP improves cement dispersion leading to higher chemical activity. Once the SP
dosage exceeds the threshold, SP starts slowing down the chemical activity of hydration
reactions. This threshold dosage was found to increase with increasing cement content
since more cement particles need to be dispersed. However, at very high cement content,
a lower SP dosage may be required since a higher percentage of cement grains acts as a
filler rather than a reacting material, as shown in Fig. 10-8.
Chapter 10 334
-350
-300
-250
-200
-150
-100
-50
00.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Superplasticizer content (%)
Aut
ogen
ous
shrin
kage
(µε)
CC=450 kg/m CC=500 kg/m CC=600 kg/m CC=750 kg/m
3
3
3
3
Figure 10-8: Effect of superplasticizer dosage on autogenous shrinkage of concrete
measured at T=20°C and w/c=0.26.
10.6.2.3 Effect of Curing Temperature
The ANN model was used to predict the autogenous shrinkage curves of concrete
mixtures with a w/c = 0.26 that were cured at various temperatures ranging from 10°C to
40°C. These curves were plotted versus the age up to 72 hours in Fig. 10-9. It is observed
that the initial slope of these curves increased with increasing curing temperature,
indicating a higher rate of autogenous shrinkage development. Moreover, the higher the
curing temperature, the earlier the curve started to flatten. At that flatten point (known as
the knee-point) a significant reduction in autogenous shrinkage occurred (Mounanga et
al., 2006). This reduction in autogenous shrinkage is frequently ascribed to the formation
of a load-bearing microstructure (self-supporting skeleton), which resists contracting
forces (Sellevold et al., 1994). Hence, increasing the curing temperature had two
Chapter 10 335
compensating influences: accelerated hydration reactions leading to a higher rate of
autogenous shrinkage, but at the same time accelerated the development of a self-
supporting skeleton that resisted autogenous shrinkage. The autogenous shrinkage is the
net resultant of these two compensating processes. Therefore, as the curing temperature
increased, a delay in starting autogenous shrinkage measurements will result in a
significant underestimation of its real value.
-600
-500
-400
-300
-200
-100
00 12 24 36 48 60
Age (hours)
Aut
ogen
ous
shrin
kage
(µε)
72
T=10°C T=15°CT=20°C T=25°CT=30°C T=35°CT=40°C
Figure 10-9: autogenous shrinkage vs. age for concrete w/c=0.26 and cured at
various temperature ranges from 10 to 40°C.
10.7. CONCLUSIONS
Autogenous shrinkage is a highly complex mechanism that makes modeling its behaviour
a difficult task. This study aimed at demonstrating the possibility of adapting artificial
neural networks to predict the autogenous shrinkage of concrete as a function of its
mixture design under different curing temperatures. A database for autogenous shrinkage
Chapter 10 336
was developed and used for training, validating and testing the ANN model. The results
support the following conclusions:
1) The ANN model thus developed is a viable method for predicating autogenous
shrinkage strains of concrete. It showed an excellent ability to capture the
interrelationships amongst key system variables.
2) Using the generalization capabilities of the ANN, the effects of several parameters
on the early-age autogenous shrinkage development could be quantified.
3) At a given age, the magnitude of autogenous shrinkage increased linearly as the
w/c was reduced. Below a w/c = 0.24, the rate of autogenous shrinkage became
more sensitive to w/c reduction. A polynomial relationship can capture the
autogenous shrinkage behaviour for w/c values ranging from 0.16 to 0.42.
4) A superplasticizer threshold dosage has been proposed at which the effect of the
SP on autogenous shrinkage switches from a motivator to chemical hydration
reactions to a deterrent. This threshold dosage changes according to the mixture
proportions.
5) The higher the curing temperature, the earlier the knee-point was reached where
the shrinkage curve started to flatten.
6) Hence, particularly at high curing temperatures, measuring autogenous shrinkage
should start as early as possible in order to capture its real value.
7) The proposed ANN is not capable of extrapolation beyond the domain of the data
used in its training. However, it can be extended beyond the current domain and
to include other experimental variables should sufficient data needed for such an
extension becomes available in the future.
Chapter 10 337
10.8. REFERENCES
ACI Committee 209, (1992), “Prediction of creep, shrinkage, and temperature effects in concrete structures (ACI 209R-92),” American Concrete Institute, Farmington Hills, MI, 47 pp.
ACI Committee 209, (2008), “Guide for modeling and calculating shrinkage and creep in
hardened concrete (ACI Report 209.2R-08),” American Concrete Institute, Farmington Hills., 44 p.
Adeli, H., (2001), “Neural networks in civil engineering: 1989-2000,” Computer-Aided
Civil and Infrastructure Engineering, Vol. 16, No. 2, pp. 126-142. Adhikary, B. and Mutsuyoshi, H., (2006), “Prediction of shear strength of steel fiber RC
beams using neural networks,” Construction and Building Materials, Vol. 20, No. 9, pp. 801-811.
Akcay, B. and Tasdemir, M., (2009), “Optimisation of using lightweight aggregates in
mitigating autogenous deformation of concrete,” Construction and Building Materials, Vol. 23, No.1, pp. 353-363.
Akkaya, Y., Ouyang, C. and Shah, S., (2007), “Effect of supplementary cementitious
materials on shrinkage and crack development in concrete,” Cement and Concrete Composites, Vol. 29, No. 2, pp.117-123.
Amegashie, F., Shang, J.Q., Yanful, E.K. Ding, W., Al-Martini, S. (2006), “Using
Complex Permittivity and Artificial Neural Networks to Identify and Classify Copper, Zinc and Lead Contamination in Soil,” Canadian Geotechnical Journal, Vol. 43, pp. 100-109
Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafaï, N., (2006),
“Autogenous deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp. 123-136.
Bažant, Z. P., (1995), “Creep and shrinkage prediction models for analysis and design of
concrete structures—Model B3,” Materials and Structures, Vol. 28, No. 6, pp. 357-365.
Bentur, A., Igarashi, S. and Kovler, K., (2001), “Prevention of autogenous shrinkage in
high-strength concrete by internal curing using wet lightweight aggregates,” Cement and Concrete Research, Vol. 31, No. 11, pp. 1587-1591.
Bentz, D. and Jensen, O., (2004), “Mitigation strategies for autogenous shrinkage
cracking,” Cement and Concrete Composites, Vol. 26, No. 6, pp. 677-685.
Chapter 10 338
Bilgehan, M. and Turgut, P., (2010), “The use of neural networks in concrete compressive strength estimation,” Computers and Concrete, Vol. 7, No. 3, pp. 271-283.
Brook, J., Cabrera, J. and Johari, M., (1999), “Factors affecting autogenous shrinkage of
silica fume high strength concrete,” In: Autogenous shrinkage of concrete: proceedings of the international workshop, Japan Concrete Institute, Tazawa E.(ed.), Taylor & Francis, pp. 195-202.
CEB-FIP 1990, (1991), “Model code for concrete structures,” Bulletin Information, No.
199, 201 pp. Daniel, C. and Ted, H., (2008), “Internal curing of high-performance concrete with pre-
soaked fine lightweight aggregate for prevention of autogenous shrinkage cracking,” Cement and Concrete Research, Vol. 38, No. 6, pp. 757-765.
Demir, F., (2008), “Prediction of elastic modulus of normal and high strength concrete
by artificial neural networks,” Construction and Building Materials, Vol. 22, No. 7, pp. 1428-1435.
Demuth, H., and Beal, M., (1998). Neural Network Tool Box for Use with MATLAB
Version 3.0, The Math Works Inc., pp. 5.37-5.39. El-Chabib, H., Nehdi, M. and Said, A., (2005), “Predicting shear capacity of NSC and
HSC slender beams without stirrups using artificial intelligence,” Computers and Concrete, Vol. 2, No. 1, pp. 79-96.
Flood, I., and Kartam, N., (1994), “Neural networks in civil engineering I: principles and
understanding,” Journal of Computing in Civil Engineering, Vol. 8, No. 2, pp. 131-148.
Gardner, N., and Lockman, M., (2001), “Design provisions for drying shrinkage and
creep of normal-strength concrete,” ACI Materials Journal, Vol. 98, No. 2, pp. 159-167.
Guangcheng, L., Youjun, X. and Zhengwut, J., (2006), “Volume changes of very-high-
performance cement-based composites,” Magazine of concrete research, Vol. 58, No. 10, pp. 657-663.
Holt, H. and Leivo, M., (2004), “Cracking risks associated with early age shrinkage,”
Cement and Concrete Composites, Vol. 26, No. 5, pp. 521-530. Holt, H., (2005), “Contribution of mixture design to chemical and
autogenous shrinkage of concrete at early ages,” Cement and Concrete Research, Vol. 35, No. 3, pp. 464-472.
Chapter 10 339
Haykin, S., Neural Networks: A Comprehensive Foundation, 2nd Edition, Upper Saddle River, NJ: Prentice-Hall, 1999, 700 p.
Igarashi, S., Bentur, A., Kovler, K., (2000), “Autogenous shrinkage and induced
restraining stresses in high-strength concretes,” Cement and Concrete Research, Vol. 30, No. 11, pp. 1701-1707.
Ince, R., (2004), “Prediction of fracture parameters of concrete by artificial neural
networks,” Engineering Fracture Mechanics , Vol. 71, No. 15, pp. 2143-2159. Konsta, m., Dattatreya, J. and Shah, S., (2003), “Influence of mineral admixtures on the
autogenous shrinkage and porosity of high strength concrete,” In: Sixth CANMET/ACI International Conference on Durability of Concrete, Malhotra V.M. (ed.), ACI-SP 212, pp. 227-238.
Lee, K., Lee, H., Lee, S. and Kim, G., (2006), “Autogenous shrinkage of concrete
containing granulated blast-furnace slag,” Cement and Concrete Research, Vol. 36, No. 7, pp. 1279-1285.
Loukili, A., Chopin, D., Khelidj, A. and Le Touzo, J., (2000), “A new approach to
determine autogenous shrinkage of mortar at an early age considering temperature history,” Cement and Concrete Research, Vol. 30, No. 6, pp. 915-922.
Lpَez, V. and Pacios, A., (2006), “Modeling the influence of SRA on properties of HPC,”
Proceedings of International symposium on Measuring, monitoring and modeling concrete properties, Konsta-Gdoutos (eds.), Springer Netherlands, Part 5, pp. 441-447.
Lura, P., Pease, B., Mazzotta, G., Rajabipour, F. and Weiss, J., (2007), “Influence of
shrinkage-reducing admixtures on development of plastic shrinkage cracks,” ACI Materials Journal, Vol. 104, No. 2, pp. 187-194.
Ma, J., Dietz, J. and Dehn, F., (2003), “Ultra high performance self-compacting
concrete,” Proceedings of 3rd International Symposium on Self-compacting Concrete, Ólafur W. and Indriði N. (eds.) , RILEM Publications, pp. 136-142.
Mak, S., Ritchie, D., Taylor, A. and Diggin, R., (1999), “Temperature effects on early
age autogenous shrinkage in high performance concretes,” In: Autogenous shrinkage of concrete: proceedings of the international workshop, Japan Concrete Institute, Tazawa E.(ed.), Taylor & Francis, pp. 155-166.
Mazloom, M., Ramezanianpour, A. and Brooks, J., (2004), “Effect of silica fume on
mechanical properties of high-strength concrete,” Cement and Concrete Composites, Vol. 26, No. 4, pp. 347-357.
Chapter 10 340
Meddah, M., Aitcin, P. and Petrov, N., (2006), “A new approach for the determination of the starting point of autogenous shrinkage strain,” Proceedings of Seventh CANMET/ACI International Conference on Durability of Concrete, Malhotra V.M. (ed.), ACI-SP 234, pp. 473-484.
Meddah, M. and Sato, R., (2010), “Effect of curing methods on autogenous shrinkage and self-induced stress of high-performance concrete,’ ACI Materials Journal, Vol. 107, No. 1, pp. 65-74.
Morin, V., Tenoudji, F., Feylessoufi, A. and Richard, P., (2001), “ Superplasticizer
effects on setting and structuration mechanisms of ultra high-performance concrete,” Cement and Concrete Research, Vol. 31, No. 1, pp. 63-71.
Mounanga, P., Baroghel-Bouny, V., Loukili, A. and Khelidj, A., (2006), “Autogenous
deformations of cement pastes: Part I. Temperature effects at early age and micro–macro correlations,” Cement and Concrete Research, Vol. 36, No. 1, pp. 110-122.
Mukherjee, A. and Biswas, S., (1997), “Artificial neural networks prediction of
mechanical behavior of concrete at high temperature,” Nuclear Engineering and Design, Vol. 178, No. 1, pp. 1-11.
Nassif, H., Suksawang, N. and Mohammed, M., (2007), “Effect of curing methods on
early-age and drying shrinkage of high-performance concrete,” Transportation Research Record: Journal of the Transportation Research Board, Vol. 1834, pp. 48-58.
Nawa, T. and Horita, T., (2004), “Autogenous shrinkage of high-performance concrete,”
Proceedings of the International Workshop on: Microstructure and Durability to Predict Service Life of Concrete Structures, Sapporo, Japan, pp. 29-38.
Nehdi, M., El-Chabib, H. and El-Naggar, M., (2001), “Predicting performance of self-
compacting concrete mixtures using artificial neural networks,” ACI Materials Journal, Vol. 98, No. 5, pp. 394-401.
Pipat, T., Toyoharu, N., Masashi, N. and Toshiki, S., (2005), “Effect of fly ash on
autogenous shrinkage,” Cement and Concrete Research, Vol. 35, No. 3, pp. 473-482.
Rattapoohm, P. and Pichai N, (2009), “An integrated approach for optimum design of
HPC mix proportion using genetic algorithm and artificial neural networks, “ Computers and Concrete, Vol. 6, No. 3, pp. 253-268.
Rafiq, M., Bugmann, G. and Easterbrook, D. “Neural network design for engineering
applications,” Computers and Structures, Vol. 79, No. 17, 2001, pp. 1541-1552.
Chapter 10 341
Sarıdemir, M., (2009), “Predicting the compressive strength of mortars containing metakaolin by artificial neural networks and fuzzy logic,” Advances in Engineering Software, Vol. 40, No. 9, pp. 920-927.
Sellevold, E., Bjøntegaard, Ø., Justnes, H. and Dahl, P., (1994), “High-performance
concrete: early volume change and cracking tendency,” Proceedings of the International RILEM Symposium on Thermal Cracking at Early Ages, Springenschmid R. (ed.), Munich, Germany, E. & F.N. Spon, London, pp. 229-236.
Shang, J., Ding, W., Rowe, R. and Josic, L., (2004), “Detecting heavy metal contamination in soil using complex permittivity and artificial neural networks,” Canadian Geotechnical Journal, Vol. 41, No. 6, pp. 1054-1067.
Sivakumar, A. and Santhanam, M., (2006), “Experimental methodology to study plastic
shrinkage cracks in high strength concrete,” Proceedings of International symposium: Measuring, monitoring and modeling concrete properties, Konsta-Gdoutos (ed.), Springer Netherlands, Part 3, pp. 291-296.
Tazawa, E. and Miyazawa, S., (1993), “Autogenous shrinkage of concrete and its
importance in concrete technology,” Proceedings of Fifth International Symposium on Creep and Shrinkage of Concrete, Bažant Z. and Carol I. (eds.), RILEM, Barcelona, pp. 159-168.
Tazawa, E., Miyazawa, S. and Kasai, T., (1995), “Chemical shrinkage and
autogenous shrinkage of hydrating cement paste,” Cement and Concrete Research, Vol. 25, No. 2, pp. 288-292.
Tazawa, E. and Miyazawa, S., (1995), “Influence of cement and admixture on
autogenous shrinkage of cement paste,” Cement and Concrete Research, Vol. 25, No. 2, pp. 281-287.
Tazawa, E., Sato, R., Sakai, E. and Miyazawa, S., (2000), “Work of JCI committee on
autogenous shrinkage” Proceedings of international RILEM workshop: Shrinkage of concrete –Shrinkage 2000, V. Baroghel-Bouny and P.-C. Aitcin (eds.), RILEM publications S.A.R.L., Paris, France, pp. 21-40.
Topçu, I and Sarıdemir, M., (2007), “Prediction of properties of waste AAC aggregate
concrete using artificial neural networks,” Computational Materials Science, Vol. 41, No. 1, pp. 117-125.
Topçu, I. and Sarıdemir, M., (2008), “Prediction of mechanical properties of recycled
aggregate concretes containing silica fume using artificial neural networks and fuzzy logic,” Computational Materials Science, Vol. 41, No. 1, pp. 74-82.
Toutanji, H., Delatte, N., Aggoun, S., Duval, R. and Danson, A., (2004), “Effect of
supplementary cementitious materials on the compressive strength and durability
Chapter 10 342
of short-term cured concrete,” Cement and Concrete Research, Vol. 34, No. 2, pp. 311-319.
Vanikar, S., (2004), “The advances and barriers in application of new concrete
technology,” Proceedings of the international workshop on sustainable development and concrete technology, Kejin Wang (ed.), Beijing, China, pp. 25-33.
Venkiteela, G, Gregori, A., Sun, Z. and Shah, S., (2010), “Artificial neural network
modeling of early-age dynamic young’s modulus of normal concrete,” ACI Materials Journal, Vol. 107, No. 3, pp. 282-290.
West, P., Brockett, P. and Golden, L., (1997), “A comparative analysis of neural
networks and statistical methods for predicting consumer choice,” Marketing Science, Vol. 16, No. 4, pp. 370-391.
Yang, Y., Sato, R. and Kawai, K., (2005), “Autogenous shrinkage of high-strength
concrete containing silica fume under drying at early ages,” Cement and Concrete Research, Vol. 35, No. 3, pp. 449-456.
Yeh, I-C., (2008), “Modeling slump of concrete with fly ash and superplasticizer,”
Computers and Concrete, Vol. 5, No. 6, pp. 559-572. Yun, L., Seong, Y., Min, K. and Jin, K., (2006), “Evaluation of a basic creep model with
respect to autogenous shrinkage,” Cement and Concrete Research, Vol. 36, No. 7, pp. 1268-1278.
Zhang, M., Tam, C. and Leow, M., (2003), “Effect of water-to-cementitious materials
ratio and silica fume on the autogenous shrinkage of concrete,” Cement and Concrete Research, Vol. 33, No. 10, pp. 1687-1694.
Chapter 11 343
CHAPTER ELEVEN
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
11.1. SUMMARY AND CONCLUSIONS
Despite the current knowledge and specifications for the early-age shrinkage of concrete
and its mitigation techniques, numerous early-age shrinkage cracking problems in various
concrete structures have been reported, indicating the existence of high shrinkage. Yet,
the issue of early-age shrinkage and how to mitigate it has not been fully resolved and
there are clear gaps between theory, research and practice.
This dissertation attempted to overcome some of these gaps through providing a
series of fundamental investigations related to volume changes in cementitious materials
at early-age and strategies used to mitigate it, taking into consideration the ambient
environmental conditions. Moreover, new strategies for producing concrete with lower
shrinkage and cracking risk although, which is also more environmentally friendly, were
proposed. These include the use of partially hydrated cementitious materials and
wollastonite microfibers as partial replacement for cement.
At the start of this research, Chapter 2 provides a comprehensive review for the
early-age properties of concrete. It was found that the reported research on the early-age
properties and shrinkage of concrete considering the exposure environmental conditions
is rather limited. The development of early-age shrinkage for cementitious materials can
be particularly sensitive to exposure conditions in the field, including temperature and
Chapter 11 344
relative humidity. Yet, existing research studies on early-age shrinkage had a limited
scope and relied mostly on constant curing conditions. It is anticipated that if research on
the early-age shrinkage of new generations of concrete (e.g. ultra-high performance
concrete (UHPC)) continue using classical testing approaches, a similar situation to that
of normal concrete would prevail in the near future; i.e. cracking, expensive rehabilitation
and accusations of unfounded science. This will defeat the purpose of UHPC that has
been introduced as an “ultimate” building material for the construction, strengthening and
rehabilitation of bridges and other transportation infrastructure. Thus, this dissertation
introduces an integrated testing approach that allows capturing the actual mechanisms
governing shrinkage of structural elements and the role of shrinkage mitigation
techniques under simulated field-like conditions.
Chapter 3 adopts a more fundamental approach based on the progress of hydration
in an attempt to capture the effect of drying conditions on the autogenous shrinkage of
ultra-high performance concrete (UHPC) at early-ages. UHPC specimens were exposed
to different temperatures, namely, 10, 20 and 40°C under a relative humidity (RH)
ranging from 40 to 80%. Results show that autogenous and drying shrinkage are
dependent phenomena. Assuming the validity of the conventional superposition principle
between drying and autogenous shrinkage led to overestimating the actual autogenous
shrinkage under drying conditions; the level of overestimation increased with decreasing
RH. Therefore, the behaviour of autogenous strains under sealed conditions will differ
from that under drying conditions.
The findings from Chapter 3 motivated research on investigating the role and
efficiency of different shrinkage mitigation techniques under field-like conditions. For
Chapter 11 345
this purpose, new UHPC mixtures incorporating a shrinkage-reducing admixture (SRA)
and a superabsorbent polymer (SAP) as shrinkage mitigation methods were also
investigated in Chapter 4. Results show that the SRA and SAP effectiveness in reducing
autogenous shrinkage under sealed conditions differ than that under drying conditions.
Adding SRA effectively reduced drying strains, which are the dominant strains in UHPC
specimens under low RH conditions. At higher RH conditions, SRA reduced autogenous
strains, which in turn are the dominant strains in high RH environments. Under sealed
conditions, early-age expansion of SAP mixtures had a significant effect in reducing the
net strains. Under drying conditions, adding SAP resulted in higher drying strains, which
disturbed the curing process and diminished the effect of SAP as a shrinkage mitigating
method. Generally, adequate external curing is essential to mitigate early-age
deformation in UHPC even when internal curing mechanisms are provided, since it
guarantees a suitable environment for shrinkage mitigation methods to work properly.
Chapters 3 and 4 provided fundamental knowledge for the interaction
mechanisms between drying and autogenous shrinkage under various curing conditions
that simulate field-like conditions, including arid, normal and cold conditions. The
second level of the fundamental investigation covered the behaviour under the effect of
outdoor environmental conditions in which structures are subjected to moisture cycles
and/or submerged conditions (simulating submerged parts of marine and offshore
structures). This was done in chapter 5 by continuing research on the early-age shrinkage
along with applying environmental loading, including drying/wetting cycles and
submerged condition. Results show that applying the environmental load alters the
shrinkage behavior of UHPC with and without shrinkage mitigation methods. It was
Chapter 11 346
discovered that exposure to drying/wetting cycles and submerged conditions jeopardize
the effectiveness of using SRA through a washing out mechanism. Using a combination
of SRA and SAP creates a synergistic effect whereby the benefits of these shrinkage
mitigation methods are optimized. This allows overcoming their individual deficiencies
under different exposure conditions, which is promising for developing a new generation
of high-performance shrinkage mitigation admixture with a dual effect. It is concluded
that adequately considering in-situ conditions in testing protocols of UHPC should allow
gaining a better understanding of shrinkage mitigation mechanisms and developing
suitable performance specifications before a wide implementation of UHPC in full-scale
field constructions.
Fundamental investigations described in this dissertation proved that relying on
constant and standard curing conditions alone to evaluate the early-age shrinkage
behaviour and efficiency of shrinkage mitigation techniques is risky since ignoring the
realistic field-like exposure conditions may yield misleading conclusions. Obviously,
concerted efforts from concrete researchers, professional organizations and
standardization agencies are needed. In addition to the current standard tests, much
research should be conducted on developing and standardizing new series of early-age
shrinkage performance tests that concomitantly take other variables (temperature, relative
humidity, wetting-drying, immersion, aggressive media, etc.) into consideration. These
combined performance tests are better able to capture synergistic actions of chemical and
physical mechanisms in concrete structures, which can occur under real field-like
exposure conditions. Such an integrated testing approach is essential for the development
of consistent performance-based standards and specifications for concrete. Also, it should
Chapter 11 347
lead to producing more reliable data and knowledge on the early-age shrinkage behaviour
of normal and emerging concrete types with inevitable improvements in the modelling of
concrete structures.
The second core theme of this research is to achieve concrete with lower early-age
shrinkage and reduced cracking risk, along with reducing its environmental and economic
impact. UHPC has a high carbon-footprint and environmental impact due to its high
cement content, leading to high energy consumption and CO2 emissions associated with
cement production. In chapter 6, a new concept for reducing early-age shrinkage in
hardening cement-based materials is introduced. The concept consists of adding partially
hydrated cementitious materials (PHCM) as a concrete admixture. Reusing disposal
concretes in producing PHCM will yield significant economical and environmental
benefits. Results showed a high potential for PHCM to reduce concrete shrinkage. The
addition of PHCM provides hydration product micro-crystals and a pre-existing hydrated
cement paste structure, which act as passive internal restraining clusters within the fresh
concrete. Consequently, the developed passive restraining system resists bulk
deformations and acts as a load bearing structure, thus leading to lower shrinkage. Hence,
the concept of a self-restraining shrinkage system was proposed and its mechanism was
demonstrated. In addition, results show that PHCM can potentially be used for several
other purposes, including as an accelerator for concrete instead of chloride-based
accelerators that induce a risk of corrosion in reinforced concrete. A dedicated
experimental study was conducted to providence evidence for aforementioned
observation (see Appendix B).
Chapter 11 348
To complement the findings of Chapter 6, the efficiency of PHCM as a shrinkage
reducing technique, when PHCM is used separately or combined with SRA and SAP,
was investigated in Chapter 7. Results indicate that SRA was the most effective in
reducing shrinkage; however, PHCM achieved excellent early-age shrinkage reduction
when used alone or combined with other shrinkage mitigation techniques. PHCM
mitigates undesirable behaviour induced by SRA and/or SAP. Combining PHCM and
SRA mitigated the drawbacks of SRA including preventing delays in setting time and
mitigating significant reductions in early-age compressive strength. The addition of
PHCM to a mixture incorporating SAP improved the shrinkage behaviour and overcame
the SAP drawbacks including higher mass loss and porosity, which led to preventing the
reduction in early-age compressive strength induced by SAP.
Making concrete a more environmentally friendly material can be achieved in
several ways. In Chapter 6 and 7, the production of low shrinkage environmentally
friendly UHPC through using PHCM is proposed. This reduces the amount of portland
cement in the mixture by partially replacing it with other suitable more environmentally
friendly materials, thus reducing the amount of energy consumed and CO2 emitted. In a
low w/c concrete, a high fraction of cement particles remain un-hydrated due to limitation
of the available water and space. In chapter 8, the potential for using wollastonite
microfibers as partial replacement for cement by volume and it capability to reduce
shrinkage and improve the cracking resistance was investigated.
Commercially available natural wollastonite microfibers were used at three
concentrations (4, 8 and 12%) as partial substitution for cement by volume. Three
microfiber sizes were used in this study: MF1 (length 152 µm, diameter 8 µm), MF2
Chapter 11 349
(length 50 µm, diameter 5 µm), and MF3 (length 15 µm, diameter 3 µm). Results show
that the early-age properties of UHPC mixtures incorporating wollastonite microfibers
are highly affected by the microfiber aspect ratio and content, along with the hydration
age of the cementitious matrix. Wollastonite microfibers were found to reduce shrinkage
strains and increase the cracking resistance. It acts as an internal restraint for shrinkage,
reinforces the microstructure at the micro-crack level and promotes pore discontinuity,
thus leading to lower mass loss and less drying shrinkage. Incorporating wollastonite
microfibers enhanced the early-age engineering properties of the UHPC matrix along
with reducing its cement factor, which represents economic and environmental benefits.
The promising results of using wollastonite microfibers introduced in Chapter 8
encouraged the combination of wollastonite microfibers and SRA to optimize the gained
benefit. In chapter 9, the effects of incorporating SRA and/or wollastonite microfibers on
the early-age shrinkage behaviour and cracking potential of UHPC has been evaluated.
Wollastonite microfibers were added at rates of 0, 4 and 12% as partial volume
replacement for cement, while SRA was added at 1% and 2% by cement weight. Results
show that the reinforcing effect induced by wollastonite microfibers mitigated the
reduction in mechanical properties induced by SRA. Addition of wollastonite microfibers
to SRA mixtures did not impart a significant change in the measured free shrinkage
strain, while it enhanced the cracking resistance compared to that of mixtures
incorporating SRA alone. Moreover, adding wollastonite microfibers promoted pore
discontinuity, thus reducing the washing out of SRA from concrete under submerged
conditions. Consequently, the efficiency of SRA in reducing shrinkage under submerged
conditions was improved owing to the presence of wollastonite microfibers.
Chapter 11 350
In Chapter 10, an original approach based on artificial neural networks (ANN)
was proposed to assist engineers and quality control personnel in developing adaptable
inference systems for the early-age shrinkage assessment of concrete. This is practically
effective to select optimum concrete mixtures for specific exposure conditions. The ANN
model inputs include the hydration age, water-to-cement ratio, total aggregate volume,
cement content, curing temperature, and type and percentage of supplementary
cementitious materials, while the autogenous shrinkage of concrete was the single model
output. In such an ANN system, combining the effects of concrete mixture proportions
and performance criteria with respect to a certain field-like exposure condition could
simplify the decision-making process and improve the reliability of assessment.
The ANN model developed in Chapter 10 was trained using a newly assembled
database. These data can be considered as an enhancement for the existed concrete
shrinkage database, since it included testing results for modern concretes under various
environmental conditions. The ANN model showed high capability to accurately predict
the autogenous shrinkage behaviour under different exposure conditions and exhibited a
good generalization capacity beyond the training stage as validated by results obtained on
new testing data within the range of training database.
This model could be used in the design and prequalification stage of a
construction project to reduce the need for exhaustive trial batches and experimental
programs, thus facilitating the decision-making process. The developed ANN model is
versatile and can be re-trained to encompass wider ranges of input variables, different
mixture proportions, other environmental conditions, etc., when such data becomes
available.
Chapter 11 351
Among the model inputs, the w/c, superplasticizer content and curing temperature
were selected due to limited and/or contradictory results about their effect on autogenous
shrinkage. Thus, the generalization capabilities of the ANN model were used to resolve
this issue and clarify the influences of these parameters on the early-age autogenous
shrinkage development.
11.2. RECOMMENDATIONS FOR FUTURE WORK
1. Based on the findings of the current thesis, it is suggested that other early-age
properties of cement-based materials, such as their mechanical properties, be re-
examined using an integrated testing approach in which chemical, physical and
structural aspects are assessed under a field-like exposure. Such an approach can
elucidate synergistic actions of multiple mechanisms and better capture performance
risks of cement-based materials that might be overlooked by current single-parameter
standard test methods.
2. On-site cracking of concrete is usually attributed to drying shrinkage and evaporation
of water from the concrete surface. However, mechanisms and causes of cracking
differ depending on the exposure conditions, type of the concrete element, etc. It is
recommended that field tests be conducted to better quantify the actual source of
cracking stresses and to quantify the benefits imparted by different shrinkage
mitigation techniques.
3. Tests conducted under drying/wetting cycles and submerged conditions introduced in
the current thesis used a single curing temperature and a regular water immerging
Chapter 11 352
solution. Following a similar integrated testing approach, further research is needed
on UHPC using different curing temperatures and more aggressive immersing
solutions.
4. Given further research, the integrated testing procedures developed in this dissertation
can be improved and refined. Investigations of reproducibility, inter- and multi-
laboratory errors should also be conducted for such combined performance tests.
5. To develop shrinkage performance-based standards and specifications for concrete,
there is an essential need to improve the shrinkage measuring techniques. Moreover,
standardizing multi-factor performance tests could capture synergistic shrinkage
mechanisms that can not be identified by current testing procedures which consider
only one single factor at a time. Thus, it is recommended that standardization
agencies such as CSA and ASTM motivate research programs in this theme in the
near future.
6. It was shown that the PHCM can be a very effective technique for reducing concrete
shrinkage. However, to become a more versatile technique, further research and
development are required. This includes: (i) investigating the effectiveness of the
PHCM technique with different mixture proportions including various types of
binders, w/c values and chemical admixtures; (ii) investigating the effect of adding
PHCM as a paste without aggregate, which should reduce the volume of the added
PHCM, making it a more practical tool for precast-plants, (iii) developing a mixture
design procedure to include the PHCM, and (iv) investigating the long-term
properties of concrete incorporating the PHCM.
Chapter 11 353
7. The feasibility of implementing wollastonite microfibers in UHPC to resist shrinkage
cracking has been proven in the current dissertation. However, to become an effective
and applicable technique, further research and development are required. This
includes investigating: (i) the effectiveness of the wollastonite microfibers in UHPC
incorporating steel microfibers; (ii) the hybrid effect of adding wollastonite micro and
microfibers. (iii) and the long-term properties of concrete incorporating wollastonite
microfibers.
8. This research demonstrated the promising application of artificial intelligence in
modeling the complex early-age shrinkage behaviour of concrete. Such a versatile
modeling tool should be established based on reliable databases derived from
integrated performance tests. Indeed, more information on the early-age behaviour of
concrete with various materials proportions combined with environmental conditions
and/or structural loading is still needed. Once this information becomes available, the
artificial intelligence inference system can be further developed and be commercially
used to facilitate the decision-making process in the design/prequalification stage of
construction projects.
354
APPENDIX A
-120
-100
-80
-60
-40
-20
00 6 12 18 24
Time (hours)St
rain
(mic
ro)
10
15
20
25
30
35
Tem
pera
ture
(°C
)
Actual temperature cycleDesigned temperature cycle
Average sample strain
Figure A-1: Detailed temperature and total strain measured in concrete samples over 24 hours.
2
4
6
8
10
12
0 10 20 30 40Time (hours)
CTE
(10-6
/°C)
50
Sample 1Sample 2Sample 3Average
Figure A-2: Time-evolution of coefficient of thermal expansion.
355
APPENDIX B
SELF-ACCELERATED CONCRETE USING PARTIALLY
HYDRATED CEMENTITIOUS MATERIALS
B.1. INTRODUCTION
This study pioneers the concept of self-accelerated concrete. The effect of adding
partially hydrated cementitious materials (premade or from returned/unused concrete) on
the setting and hardening process of concrete cured at various temperatures was
investigated. The partially hydrated cementitious materials (PHCM) were added at rates
of 25, 33 and 50% of the overall batch weight. Similar mixtures incorporating chloride-
(CA) and non-chloride- (NCA) based accelerating admixtures were also tested for
comparison. The results indicate that the added PHCM alert the hydration kinetics and act
as a setting and hardening accelerator. Mixtures incorporating PHCM showed
comparable to and/or higher early-age compressive strength than that of both the control
and mixtures incorporating accelerating admixtures. Therefore, using PHCM paves the
way for self-accelerated concrete, without the need for accelerating admixtures, providing
a safe and cost effective method for precast and cast in place concrete. Using left-over
and unused concrete in this process enhances the sustainability of concrete and minimizes
disposal in ready mixed concrete operations.
A version of this Appendix has been Accepted in the ACI Materials Journal.
Appendix B 356
B.2. EXPERIMENTAL PROGRAM
This experimental program aims to produce a self-accelerated concrete without need for
accelerating admixtures. This Appendix was devoted to validating the addition of PHCM
as an accelerating technique and to evaluating its efficiency compared to that of different
accelerating admixtures. The characteristics of the tested mixtures are shown in Tables
B-1.
Table B-1 Tested mixtures
Mixture Accelerating method
PHCM (% of the batch weight)
Admixture Dosage (% of cement mass)
MC ‐‐‐‐ 0.0 ‐‐‐‐ MP1 25.0 ‐‐‐‐ MP2 33.0 ‐‐‐‐ MP3
PHCM 50.0 ‐‐‐‐
MAC1 ‐‐‐‐ 2.0 MAC2
Chloride Admixture ‐‐‐‐ 4.0
MANC1 ‐‐‐‐ 1.5 MANC2
Non‐Chloride Admixture ‐‐‐‐ 3.0
B.3. RESULTS AND DISCUSSION
B.3.1. Setting Time and Compressive Strength
B.3.1.1 PHCM technique
Regardless of the added amount of PHCM, the PHCM technique significantly enhanced
the development of early-age compressive strength. All PHCM mixtures consistently
produced higher early-age compressive strengths compared to that of MC cured at 10 and
20°C, as shown in Fig. B-1(a,b). The increase in compressive strength was directly
proportional to the added amount of PHCM. For instance, the 24 hrs compressive
strength at 10 and 20°C increased by 70% and 30% for MP1, 85% and 40% for MP2 and
Appendix B 357
by 140% and 45% for MP3 of their respective MC values. At 10°C, PHCM addition
induced higher increase in compressive strength compared to that at 20°C. This is likely
due to the acceleration effect induced by PHCM addition. Hence, the slow rate of
hydration and strength development at the low temperature of 10°C were compensated
for, leading to more hydration products and stronger microstructure (ACI Committee E-
701, 2001). Compared with the MC setting time, the PHCM technique also reduced the
setting time significantly as shown in Table B-2. These results demonstrate that the
PHCM technique can be considered as an effective setting and hardening accelerating
method.
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th (M
Pa) MC
MP1MP2MP3
T=10°C (a)
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Com
pres
sive s
treng
th (M
Pa) MC
MP1MP2MP3
(b)T=20°C
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th (M
Pa) MC
MAC1MAC2MANC1MANC2
(c) T=10°C
0
10
20
30
40
50
6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th (M
Pa) MC
MAC1MAC2MANC1MANC2
(d) T=20°C
Figure B-1: Early-age compressive strength of UHPC mixtures incorporating PHCM cured at a) 10°C and b) 20°C and UHPC mixtures incorporating CA and
NCA cured at c) 10°C and d) 20°C. [Maximum COV (10°C, 20°C): MC (0.9%, 2.0%), MP1 (1.9%, 4.8%), MP2 (2.3%, 5.6%), MP3 (2.7%, 2.2%), MAC1 (1.7%, 0.9%), MAC2
(4.2%, 4.7%), MANC1 (4.9%, 0.9%), MANC2 (3.8%, 6.8%)]. Table B-2: Initial and final setting times for different mixtures.
Appendix B 358
Temperature °C 10 20
Setting time (min)Mixture Initial Final Initial Final
MC 510 715 460 520 MP1 450 495 340 420 MP2 405 465 220 270 MP3 245 280 170 210 MAC1 425 475 240 285 MAC2 160 205 150 190 MANC1 475 540 350 395 MANC2 240 300 260 295
The improvement in compressive strength and reduction in setting time induced
by PHCM addition cannot be solely attributed to the effect of the older material. For
instance, for mixture MP3 after 6 hours of curing at 10°C, 50% of the mixed
cementitious material is 6 hours older. Hence, 50% of the cementitious material has an
age of 12 hours, while the other 50% has an age of 6 hours. Summing up 50% of the MC
compressive strength at age 6 hours (0.0 MPa) and 12 hours (1.68 MPa) results in a
compressive strength of about (0.84 MPa), which is much lower than that of MP3 at 6
hours (3.90 MPa). On the other hand, the setting time of the normal paste starts at about
510 min at 10°C, hence, adding 50% of 6-hour older paste material should theoretically
reduce the setting to about 150 min. However, the setting time was only reduced to about
245 min (which represents about 52% reduction with respect to the MC value). The
longer setting time for MP3 than the theoretical value (150 min) can be attributed to the
effect of remixing on the already developed microstructure. Remixing breaks down the
formed connections between hydration products (Lea, 1988), hence, the MP3 paste will
initially have a lower stiffness than that of the paste mixed 6 hours earlier.
Appendix B 359
B.3.1.2 Chloride-based accelerating admixture (CA)
The compressive strength results of mixtures incorporating different dosages of CA are
shown in Fig. B-1(c,d). The higher the dosage, the higher the rate of hydration reactions,
and consequently the higher the early-age compressive strength achieved, which is in
agreement with previous results (Cheeseman and Asavapisit, 1999). As shown in Fig. B-
1(c,d), MAC2 exhibited about six times higher early-age compressive strength compared
to that of MAC1 after 6 hrs from mixing, regardless of the curing temperature. In
addition, the CA improved the compressive strength more effectively at 10°C compared
to that at 20°C. This is probably due to the higher heat liberated from the accelerated
hydration reactions, offsetting the prolonged strength gain induced by the low curing
temperature (Shideler, 1952). Adding the CA reduced the initial and final setting times
significantly (Table B-2). For instance, MAC2 exhibited initial and final setting times at
10 and 20°C more than 60% shorter than that of MC.
B.3.1.3 Non-chloride based accelerating admixture (NCA)
The compressive strength results for mixtures incorporating different dosages of the NCA
are shown in Fig. B-1(c,d). Mixture MANC1 did not show a significant improvement in
very early-age compressive strength. During the first 10 hrs, MANC1 showed a lower
compressive strength than that of MC at 20°C. However, MANC1 exhibited initial and
final setting times about 24% shorter than that of MC. MANC1 showed a different
strength gain rate at 10°C compared to that at 20°C, with a significant increase in
compressive strength (about double that of MC). On the other hand, MANC2 had higher
early compressive strength and shorter setting time compared to that of MC. For instance,
MANC2 exhibited about 136% and 36% higher compressive strength than to that of MC
Appendix B 360
after 24 hrs at 10 and 20°C, respectively. Hence, the NCA can generally be considered as
a setting accelerator, confirming results reported by others (Chikh et al., 2008), while its
effect on early-age strength gain can vary depending on the added dosage.
B.3.1.4 Comparison between different acceleration techniques
To evaluate the efficiency of the PHCM accelerating technique versus accelerating
admixtures, the differences between the compressive strength gain achieved by mixtures
incorporating PHCM and that incorporating accelerating admixtures were calculated, as
shown in Fig. B-2. Generally, the PHCM mixtures exhibited higher compressive strength
compared to that of MAC1 and MANC1. For instance, the differences in early-age
compressive strength between PHCM mixtures and MAC1, over the investigated period,
ranged respectively from -0.6 to +10 MPa at 10°C, and from +1 to +15 MPa at 20°C,
respectively. A similar trend was observed between PHCM mixtures and MANC1, as
shown in Fig. B-2(c).
On the other hand, mixtures incorporating PHCM showed lower early-age
compressive strength compared to that of MAC2 mixtures. The gap in compressive
strength decreased with increasing the added amount of PHCM. However, the mixture
MP3 incorporating 50% PHCM showed early-age compressive strength comparable to
and/or higher than that of MANC2 at 10 and 20°C over the investigated period.
Moreover, MP3 exhibited a compressive strength at the end of the investigated period
higher than that of MAC2 as shown in Fig. B-2(b). In simple terms, it takes the
maximum dosage of CA to compete with the acceleration induced by the PHCM method.
Appendix B 361
-16
-12
-8
-4
0
4
8
12
16
6 8 10 12 24 6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th dif
feren
ce(M
Pa)
MP1 Vs. MAC1MP2 Vs. MAC1MP3 Vs. MAC1
T=20°CT=10°C(a)
-16
-12
-8
-4
0
4
8
12
16
6 8 10 12 24 6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th dif
feren
ce (M
Pa)
MP1 Vs. MAC2MP2 Vs. MAC2MP3 Vs. MAC2
T=20°CT=10°C(b)
-16
-12
-8
-4
0
4
8
12
16
6 8 10 12 24 6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th dif
feren
ce (M
Pa)
MP1 Vs. MANC1MP2 Vs. MANC1MP3 Vs. MANC1
T=20°CT=10°C(c)
-16
-12
-8
-4
0
4
8
12
16
6 8 10 12 24 6 8 10 12 24Age (hours)
Comp
ress
ive st
reng
th dif
feren
ce (M
Pa)
MP1 Vs. MANC2MP2 Vs. MANC2MP3 Vs. MANC2
T=20°C(d) T=10°C
Figure B-2: Compressive strength difference between UHPC mixtures incorporating PHCM and that incorporating CA [ a) MAC1, and b) MAC2] and NCA [ c) MANC1,
and d) MANC2 ] at different curing temperatures.
Figure B-3 shows the compressive strength results for different mixtures at age
28-days. Generally, all mixtures incorporating different acceleration techniques exhibited
28-days compressive strength comparable to that of MC, regardless of the curing
temperature. For instance, the difference in 28-days strength between the control mixture
without accelerating technique and mixtures with accelerating techniques at 10°C and
20°C, ranged from -1% to +6.5% and from -4% to +2.5% with respect to the MC value,
respectively. This slight variation in the 28-days compressive strength can be attributed to
the fact that at low w/c ratio, the hydration progress is mainly controlled by the
availability of sufficient space for hydration products to form (Lea, 1988).
Appendix B 362
0
30
60
90
120
150
180
10 20Curing Temperature (°C)
Comp
ress
ive st
reng
th (M
Pa)
MC MAC1MAC2 MANC1MANC2 MP1MP2 MP3
Figure B-3: 28 days compressive strength of UHPC mixtures incorporating PHCM , CA and NCA cured at 10°C and 20°C.
The effectiveness of the PHCM technique as a setting accelerator increased with
increasing the added amount of PHCM, as shown in Fig. B-4. The lower the temperature,
the higher the relative acceleration induced by the PHCM technique. Mixtures
incorporating PHCM achieved longer setting time compared to that of MAC2 (see Fig.
B-4(b)). However, MP2 and MP3 exhibited similar and/or shorter setting time compared
to MAC1 (see Fig. B-4(a)). On the other hand, PHCM mixtures exhibited shorter setting
time compared to that of MANC1, while longer compared to that of MANC2, except for
MP3 (see Fig. B-4(c,d)).
Appendix B 363
-200
-150
-100
-50
0
50
100
150
200
Initial Final Initial Final
Settin
g tim
e cha
nge (
%)
MP1 Vs. MAC1MP2 Vs. MAC1MP3 Vs. MAC1(a)
T=10°C T=20°C
-200
-160
-120
-80
-40
0
40
80
120
160
200
Initial Final Initial Final
Settin
g tim
e cha
nge (
%)
MP1 Vs. MAC2MP2 Vs. MAC2MP3 Vs. MAC2(b)
T=10°C T=20°C
-200
-160
-120
-80
-40
0
40
80
120
160
200
Initial Final Initial Final
Settin
g tim
e cha
nge (
%)
MP1 Vs. MANC1MP2 Vs. MANC1MP3 Vs. MANC1(c)
T=10°C T=20°C
-200
-160
-120
-80
-40
0
40
80
120
160
200
Initial Final Initial Final
Settin
g tim
e cha
nge (
%)
MP1 Vs. MANC2MP2 Vs. MANC2MP3 Vs. MANC2(d)
T=10°C T=20°C
Figure B-4: Change in initial and final setting times for UHPC mixtures incorporating PHCM and that incorporating CA [ a) MAC1, and b) MAC2] and
NCA [ c) MANC1, and d) MANC2 ] at different curing temperatures.
B.3.2. Degree of Hydration
Early-age strength development of concrete mixtures highly relies on the degree of
hydration achieved (Xiao and Li, 2008). The correlation between the compressive
strength and degree of hydration (represented by the amount of BW) is plotted in Fig. B-
5. It can be observed that the relationship exhibits a linear trend for mixtures
incorporating PHCM and CA (with R2 = 0.97 and 0.94, respectively). Conversely,
combining data points for MANC1 and MANC2 mixtures indicates a very poor linear
trend (R2 = 0.67), while separating these two sets of data leads to a linear trend with R2 =
0.97 and 0.95 for MANC1 and MANC2, respectively, as shown in the upper left corner of
Appendix B 364
Fig. B-5. This linear relationship between compressive strength and degree of hydration
has been reported in previous works (Xiao and Li, 2008, Powers, 1948). It indicates a
proportional relationship between the added dosages of PHCM and CA and the
development of hydration and strength. Conversely, the NCA behaves differently from
one dosage to another, in compliance with previous studies (Rear and Chin, 1990).
R2 = 0.97
R2 = 0.94R2 = 0.67
0
10
20
30
40
50
0 0.02 0.04 0.06 0.08 0.1 0.12Amount of BW (mg/mg binder)
Comp
ress
ive st
reng
th (M
Pa)
MP1MP2MP3MAC1MAC2MANC1MANC2MP
R2 = 0.97
R2 = 0.95
0
10
20
30
40
50
0 0.025 0.05 0.075 0.1Amount of BW (mg/mg binder)
Comp
ress
ive st
reng
th (M
Pa)
01000200030004000500060007000
Comp
ress
ive st
reng
th (p
si)
MANC1MANC2Series11
Figure B-5: Relation between degree of hydration (amount of BW) and compressive
strength development for the tested mixtures.
B.3.3. Heat of Hydration
Mixtures incorporating PHCM had similar temperature evolution curves that differed
from that of the control MC, as shown in Fig. B-6(a), indicating a variation in the
hydration kinetics. Initially, the temperature for PHCM mixtures rised rapidly after
casting the specimen in the semi-adiabatic cell (20 min from water addition) and did not
exhibit an induction period (period in which the rate of hydration reactions slows down
significantly (Ramachandran et al., 2002). Temperature rised up until reaching a peak of
about 42°C after about 7.5 hrs for MP3, 38°C and 39°C after about 8 hrs for MP1 and
Appendix B 365
MP2, respectively. On the other hand, MC hydration exhibited induction and acceleration
periods as shown in Fig. B-6(a). The acceleration period for MC was initiated at about 6
hrs later than that of PHCM mixtures and had a temperature peak of about 36°C at around
12 hrs from water addition. Hence, the PHCM technique effectively diminished the
induction period, leading to a continuous progress of hydration reactions and
consequently shorter setting time and higher early-age compressive strength as discussed
earlier.
20
25
30
35
40
45
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (hrs)
Temp
eratu
re (°
C)
MCMP1MP2MP3
(a)
20
25
30
35
40
45
0 2 4 6 8 10 12 14 16 18 20 22 24Time (hrs)
Temp
eratu
re (°
C)
MCMAC1MAC2
(b)
20
25
30
35
40
45
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (hrs)
Temp
eratu
re (°
C)
MANC1MANC2
(c)
Figure B-6: Heat of hydration for UHPC mixtures incorporating a) PHCM, b) CA, and c) NCA.
The general profile of temperature evolution curves for mixtures incorporating CA
was similar to that of MC but with different rates. In general, as the dosage of the
Appendix B 366
admixture increased, hydration reactions were accelerated. This is apparent in the heat of
hydration curves in Fig. B-6(b). With increased admixture dosage, the curves shifted
upward and to the left, which is in agreement with previous work (Cheeseman and
Asavapisit, 1999, Ramachandran et al., 2002).
For mixtures incorporating the NCA, a similar trend of temperature evolution
curves to that of the PHCM mixture was observed, as shown in Fig. B-6(c). The dormant
period was overlapped by the acceleration period, in agreement with previous results
(Hill and Daughert, 1996). However, mixtures incorporating the NCA showed lower
temperature peaks compared to that of the PHCM mixtures. This suggests higher
reactivity at early-age for the PHCM mixtures with respect to that incorporating the NCA.
B.3.4. Shrinkage
Shrinkage results for mixtures incorporating PHCM and accelerating admixtures are
shown in Fig. B-7. All mixtures incorporating accelerating admixtures showed a
significant increase in the measured shrinkage (about 31, 10, 12 and 23% for MAC1,
MAC2, MANC1 and MANC2, respectively), while no significant mass change was
measured, which is in agreement with previous works (Shideler, 1952, Rixom, M.R. and
Mailvaganam, 1999) (Fig. B-8). Conversely, PHCM mixtures showed lower shrinkage
and mass loss compared to that of the control MC (Figs. B-7 and B-8). This can be
attributed to the fact that the measured shrinkage includes drying and autogenous
shrinkage (thermal deformation can be ignored due to the small cross-section of the
samples (Baroghel-Bouny et al., 2006). Autogenous shrinkage is strongly related to
hydration reactions in which water is consumed, leading to internal self-desiccation
Appendix B 367
(without external water loss) (Powers, 1948). For low w/c mixtures, the amount of
autogenous shrinkage can be comparable to that of drying shrinkage (Tazawa, E. and
Miyazawa, 1999). Using accelerating admixtures increases the rate of hydration and
consequently refines and changes the size distribution of capillary pores, which in turn
eliminates water loss (ACI Committee 301, 2005). Conversely, the accelerated hydration
reactions and finer porosity will increase self-desiccation, and consequently increase the
autogenous shrinkage contribution to the measured shrinkage.
-800
-700
-600
-500
-400
-300
-200
-100
00 1 2 3 4 5 6
Time (day)
Shrin
kage
(µε
)
7
MCMP1MP2MP3
(a)
-800
-700
-600
-500
-400
-300
-200
-100
00 1 2 3 4 5 6 7
Time (day)
Shrin
kage
(µε
)
MCMAC1MAC2MANC1MANC2
(b)
Figure B-7: Drying shrinkage for a) PHCM mixtures and b) mixtures incorporating accelerating admixtures [Maximum COV: MC (4.5%), MAC1 (4.8%), MAC2 (1.9%),
MANC1 (5.9%), MANC2 (4.3%), MP1 (3.1%), MP2 (1.5%), MP3 (4.6%)].
-1.0
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.00 1 2 3 4 5 6 7
Time (day)
Mass
loss
(%)
MCMP1MP2MP3
(a)
-1.0
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.00 1 2 3 4 5 6 7
Time (day)
Mass
loss
(%)
MCMAC1MAC2MANC1MANC2
(b)
Figure B-8: Mass loss for a) PHCM mixtures, and b) mixtures incorporating accelerating admixtures.
Appendix B 368
In PHCM mixtures, a similar process takes place; however, the added PHCM can
act as a passive internal restraint, thus reducing the amount of deformation developed,
along with accelerating the hydration reactions. At the micro-level, the added PHCM
provides CH crystals that can act as a passive restraint and reduce the measured physical
shrinkage (Jensen and Hansen, 1996). The hypothesis that CH crystals can provide
restraint was proposed by (Bentz and Jensen, 2004) based on results reported in (Carde
and Francois, 1997, Powers, 1962) which indicate a significant effect of leaching and
dissolution of CH crystals on mechanical and deformation properties. This explanation is
consistent with autogenous shrinkage results shown in Fig. B-9, and with DSC results
presented below. Increasing the amount of added PHCM increased the passive internal
restrain and consequently reduced the amount of autogenous shrinkage. Accelerating
admixtures initially increase the rate of autogenous shrinkage, as a result of increasing the
rate of hydration, until reaching adequate stiffness to withstand shrinkage as reported by
(Anna et al., 1955).
-900
-800
-700
-600
-500
-400
-300
-200
-100
00 12 24 36 48 60 72 84 9
Time (hours)
Autog
enou
s shr
inkag
e (µε
)
6
MC MP1MP2 MP3MAC2 MANC2
Figure B-9: Autogenous shrinkage for a) PHCM mixtures and b) mixtures incorporating accelerating admixtures.
Appendix B 369
Furthermore, the evaporable water content in PHCM mixtures is probably less
than that in the control MC and mixtures incorporating accelerating admixtures at the
onset of drying. This can be explained as follows: all mixtures initially have the same
amount of mixing water, however, the added mixing water before casting the specimens
for PHCM mixtures (i.e. mixing water at the second stage) is less than that of the
ordinary mixture. For instance, for the mixture MP3, half of the mixing water is added in
the first mixing stage, while the other half is added at the second stage before casting the
shrinkage specimens. Hence, PHCM mixtures possess lower evaporable water and
consequently lower mass loss as shown in Fig. B-8.
B.4. CONCLUSIONS
The main conclusions that can be drawn from this experimental investigation are the
following:
1) The addition of PHCM had a strong acceleration effect on the early-age setting
and hardening process for concrete, hence, it can be considered as a setting and
hardening accelerator.
2) Mixtures incorporating PHCM showed comparable and/or better early-age
compressive strength results than that of the control and mixtures incorporating
accelerating admixtures at cold (10°) and normal (20°C) temperatures.
3) The higher the added portion of PHCM, the higher was the acceleration effect.
4) Mixtures incorporating PHCM achieved a high potential for reducing autogenous
shrinkage through providing internal passive restraint, compared to mixtures
Appendix B 370
incorporating accelerating admixtures. This is a clear advantage compared to
conventional accelerating admixtures, which generally increase shrinkage strains.
5) The addition of PHCM mainly affects the nucleation and renewal process of CSH.
6) The PHCM technique pioneers the concept of self-accelerated concrete, which
has a paramount potential, particularly in the pre-cast industry. It resolves the two
major drawbacks associated with chloride-based accelerators; corrosion related to
chlorides and increased shrinkage.
7) Concrete sustainability can be enhanced through using left-over and returned
concrete in producing self-accelerated concrete, thus preventing wastage and
disposal costs.
Appendix B 371
B.5. REFERENCES
ACI Committee E-701, (2001) Cementitious Materials for Concrete, ACI Education Bulletin E3-01, American Concrete Institute, Farmington Hills, Michigan, 25 p.
ACI Committee 301, (2005) Specifications for Structural Concrete (ACI 301-05),”
American Concrete Institute, Farmington Hills, MI, 49 p. Anna, K., Leivo, M. and Sipari, P., (1995), “Experimental study on the basic phenomena
of shrinkage and cracking of fresh mortar,” Cement and Concrete Research, Vol. 25, No. 8, pp. 1747-1754.
Baroghel-Bouny, V., Mounanga, P., Khelidj, A., Loukili, A. and Rafaï, N., (2006),
“Autogenous deformations of cement pastes: Part II. W/C effects, micro–macro correlations, and threshold values,” Cement and Concrete Research, Vol. 36, No. 1, pp. 123-136.
Bentz, D.P. and Jensen, O.M., (2004), “Mitigation strategies for autogenous shrinkage
cracking,” Cement and Concrete Composites, V. 26, No. 6, pp. 677-685. Carde, C. and Francois, F., (1997), “Effect of the leaching of calcium hydroxide from
cement paste on mechanical and physical properties,” Cement and Concrete Research, Vol. 27, No. 4, pp. 539-550.
Cheeseman, C.R. and Asavapisit, S., (1999), “Effect of calcium chloride on the hydration
and leaching of lead-retarded cement,” Cement and Concrete Research, Vol. 29, No. 6, pp. 885-892.
Chikh, N., Cheikh-Zouaoui, M., Aggoun, S. and Duval, R., (2008), “Effects of calcium
nitrate and triisopropanolamine on the setting and strength evolution of Portland cement pastes,” Materials and Structures, Vol. 41, No. 1, pp. 31-36.
Hill, R. and Daughert, K., (1996), “The interaction of calcium nitrate and a class C fly
ash during hydration,” Cement and Concrete Research, V. 26, No. 7, pp. 1131-1143.
Jensen, O.M. and Hansen, P.F., (1996), “Autogenous deformation and change of relative
humidity in silica fume modified cement paste,” ACI Materials Journal, Vol. 93, No. 6, 1996, pp. 539-543.
Lea, F.M., (1988). Lea's Chemistry of Cement and Concrete, 4th Edition, (ed.) P.C.
Hewlett, J. Wiley, New York, 1053 p. Powers, T.C. and Brownyard, T.L., (1948), “Studies of the physical properties of
hardened portland cement paste,” Bulletin No. 22., Research Laboratories of the Portland Cement Association. Reprinted from Journal of the American Concrete Institute, October 1946–April 1947, Proceedings 43, Detroit, pp. 971-992.
Appendix B 372
Powers, T.C., (1962), “A hypothesis on carbonation shrinkage,” Journal of the Portland
Cement Association Research & Development Laboratories, Vol. 4, No. 2, 1962, pp. 40-50.
Ramachandran, V.S., Haber, R.M., Beaudoin, J.J. and Delgado, A.H., (2002). Handbook
of Thermal Analysis of Construction Materials, Noyes & William Andrew Publishing, Norwich, 655 p.
Rear, K. and Chin, D. (1990), “Non-chloride accelerating admixtures for early
compressive strength,” Concrete International, Vol. 12, No. 10, pp. 55-58. Rixom, M.R. and Mailvaganam, N.P., Chemical Admixtures for Concrete, 3rd Edition,
Brunner-Routledge, New York, 1999, 431 p. Shideler, J.J., (1952), “Calcium cholride in concrete,” ACI Materials Journal, Vol. 48,
No. 3, pp. 537-559. Tazawa, E. and Miyazawa, S., (1999), “Effect of constituents and curing condition on
autogenous shrinkage of concrete,” In: Autogenous Shrinkage of Concrete: Proceedings of the International Workshop, (ed.) Eiichi Tazawa, Taylor & Francis, pp. 269-280.
Xiao, L. and Li, Z., (2008), “Early-age hydration of fresh concrete monitored by non-
contact electrical resistivity measurement,” Cement and Concrete Research, Vol. 38, No. 3, pp. 312-319.
373
APPENDIX C
Table C -1: Sample for data used in ANN Model
W/C A/C C.C. SF FA HRWRA Temp. Age Shrinkage strain Ref.
‐‐‐‐‐ ‐‐‐‐ Kg (% )* (% )* (% )* ( °C) (Hours) (μ m/m) 0.28 2.82 576 10 0 1.8 20 5.5 ‐17.4334 0.28 2.82 576 10 0 1.8 20 6 ‐153.995 0.28 2.82 576 10 0 1.8 20 7 ‐161.743 0.28 2.82 576 10 0 1.8 20 8 ‐167.554 0.28 2.82 576 10 0 1.8 20 9 ‐175.303 0.28 2.82 576 10 0 1.8 20 10 ‐178.208 0.28 2.82 576 10 0 1.8 20 11 ‐183.051 0.4 4.14 424 10 0 1.8 20 11 ‐107.506 0.4 4.14 424 10 0 1.8 20 12 ‐106.538 0.4 4.14 424 10 0 1.8 20 13 ‐107.506 0.4 4.14 424 10 0 1.8 20 14 ‐106.538 0.4 4.14 424 10 0 1.8 20 15.574 ‐104.977 0.5 5.36 340 10 0 1.4 20 12 ‐84.2615 0.5 5.36 340 10 0 1.4 20 13 ‐84.2615 0.5 5.36 340 10 0 1.4 20 14 ‐84.2615 0.5 5.36 340 10 0 1.4 20 15 ‐87.1671 0.5 5.36 340 10 0 1.4 20 15.5185 ‐88.4956 0.5 5.36 340 10 0 1.4 20 17 ‐88.4956 0.5 5.36 340 10 0 1.4 20 18.4815 ‐90.2655
[3]
0.5 5.36 340 10 0 1.4 20 30.037 ‐99.115 0.22 2.036 750 0 0 1.4 20 1.27434 ‐20.9129 0.22 2.036 750 0 0 1.4 20 2.23009 ‐28.7028 0.22 2.036 750 0 0 1.4 20 3.13274 ‐60.1828 0.22 2.036 750 0 0 1.4 20 3.66372 ‐96.9667 0.22 2.036 750 0 0 1.4 20 4.61947 ‐146.862 0.22 2.036 750 0 0 1.4 20 5.68142 ‐178.324
[4]
0.22 2.036 750 0 0 1.4 20 7.43363 ‐214.974 - w/c: water to cement ratio - a/c: aggregate to cement ratio - CC: Cement Content - SF: Silica fume - FA: Fly Ash - HRWRA: High range water reducing admixture - * : % of the cement weight.
374
APPENDIX D
20
23
26
29
32
35
38
41
44
0 6 12 18 24 30 36 42 48
Time (hours)
Te
mp
era
ture
(°C
)C
M18
M28
M38
(b)(a)(a)
20
23
26
29
32
35
38
41
44
0 6 12 18 24 30 36 42 48
Time (hours)
Tem
pera
ture
(°C
)
C
M112
M212
M312
(c)(b)
Figure D-1: Heat of hydration for mixtures incorporating wollastonite microfibers
with a) 8% and b) 12% contents compared to that of the control mixture.
Appendix D 375
-600
-500
-400
-300
-200
-100
0
0 24 48 72 96 120 144 16
Age (hours)
To
tal
sh
rin
kag
e (
µ)
8
C
M18
M28
M38
(b)(a)
-600
-500
-400
-300
-200
-100
0
0 24 48 72 96 120 144 16
Age (hours)
To
tal
sh
rin
kag
e (
µ)
8
C
M112
M212
M312
(c)(b)
Figure D-2: Total shrinkage for mixtures incorporating wollastonite microfibers
with a) 8% and b) 12% contents compared to that of control mixture.
Appendix D 376
-600
-500
-400
-300
-200
-100
0
100
0 24 48 72 96 120 144 168
Age (hours)
Au
tog
en
ou
s s
tra
in (
µ)
C
M18
M28
M38
(b)(a)
-600
-500
-400
-300
-200
-100
0
100
0 24 48 72 96 120 144 168
Age (hours)
Au
tog
en
ou
s s
tra
in (
µ)
C
M112
M212
M312
(c)(b)
Figure D-3: Autogenous shrinkage for mixtures incorporating wollastonite
microfibers with a) 8% and 6) 12% contents compared to that of the control
mixture.
377
APPENDIX E
Table E -1: Trial Mixtures Composition
No. Cement Content (kg)
Silica Fume(kg)
Quartz Sand (kg)
Quartz Powder (kg)
Water (kg)
HRWRA (kg)
1 720 210 1020 232 110 31 2 650 165 779 560 237 23 3 800 200 871 463 150 25 4 930 104 930 451 155 24 5 820 92 798 622 155 13 6 1114 169 611 305 212 40 7 1290 350 462 90 220 35
378
VITA Name: Ahmed Mohammed Salah El-Din Mohammed Soliman Post Secondary Ain Shams University Education and Cairo, Egypt Degrees: 1997-2002 B.Sc. (Civil Engineering) Ain Shams University Cairo, Egypt 2002-2005 M.Sc. (Civil Engineering/Materials) The University of Western Ontario London, Ontario, Canada 2007-2011, Ph.D. (Civil Engineering/Materials) Related Work Research and Teaching Assistant Experience: The University of Western Ontario
2007-2011
Research and Teaching Assistant Ain Shams University 2002-2007
Quality Control, Planning and Controls Engineer
Concrete Material Unit, Ain Shams University, Egypt 2002-2007
PUBLICATIONS
1. Nehdi, M. L. and Soliman, A. M. (2009) " Early-age properties of concrete: Overview of fundamental concepts and state-of-the art research," Construction Materials, ICE, Vol. 164, Issue CM2, April 2011, pp. 55-77.
2. Soliman, A. M. and Nehdi, M. L. (2009) “Effect of drying conditions on autogenous shrinkage of ultra-high performance concrete at early-age," Materials and Structures, RILEM, In Press (published on line in 2011- DOI: 10.1617/s11527-010-9670-0).
3. Soliman, A. M. and Nehdi, M. L. (Accepted 2011) “Early-age shrinkage of ultra high- performance concrete under drying/wetting cycles and submerged conditions," ACI Materials Journal, American Concrete Institute, Accepted.
4. Soliman, A. M. and Nehdi, M. L. (Accepted 2010) “Self-accelerated reactive powder concrete using partially hydrated cementitious materials," ACI Materials Journal, American Concrete Institute, In Press (to appear in 2011).
379
5. Soliman, A. M. and Nehdi, M. L. (2010) “Self-restraining concrete: mechanisms and Evidence," Cement and Concrete Research, Elsevier Science, Submitted
6. Soliman, A. M. and Nehdi, M. L. (2010) “Can partially hydrated cementitious materials mitigate early-age shrinkage in ultra-high Performance concrete?,” Cement and Concrete Research, Elsevier Science, Submitted
7. Soliman, A. M. and Nehdi, M. L. (2010) “Influence of natural wollastonite microfiber on early-age behaviour of ultra-high performance concrete," Journal of Materials in Civil Engineering, ASCE, Submitted
8. Soliman, A. M. and Nehdi, M. L. (2011) “Shrinkage behaviour of ultra high-performance concrete with shrinkage reducing admixture and wollastonite microfiber,” Cement and Concrete Composite, Elsevier Science, Submitted.
9. Soliman, A. M. and Nehdi, M. L. (2011) “Artificial neural network modeling of early-age autogenous shrinkage of concrete,” ACI Materials Journal, American Concrete Institute, Submitted
10. Soliman, A. M. and Nehdi, M. L. (2010) “Performance of shrinkage reducing admixture under drying/wetting cycles and submerged conditions,” Proceedings of 2nd International Structures Specialty Conference, Winnipeg, Manitoba, Canada, pp. (ST-30-1: ST-30-7).Published
11. Soliman, A. M. and Nehdi, M. L. (2010) “Effect of shrinkage mitigation methods on early-age shrinkage of UHPC under simulated field conditions,” Eighth International Conference on. Short and Medium Span Bridges, Niagara Falls, Ontario, Canada, 2010, pp. (83-1 :83-8) Published
12. J. Camiletti, Soliman, A.M. and M.L. Nehdi (2010). “Performance of nano-limestone as a cement hydration accelerator” Proceedings of 2nd International Structures Specialty Conference, Winnipeg, Canada, pp. (ST-31-1 - ST-31-7). International Conference. Published (Co-supervising Work).
13. Soliman, A. M. and Nehdi, M. L. (2012) “Early-age behaviour of wollastonite microfiber UHPC” Twelfth International Conference on Recent Advances in Concrete Technology and Sustainability Issues, Prague, Czech Republic, 2012. Submitted.
14. Soliman, A. M. and Nehdi, M. L. (2012) “Performance of concrete shrinkage mitigation admixtures under simulated field conditions,” Proceedings of the Tenth International Conference on Superplasticizer and Other Chemical Admixtures in Concrete to be held in Prague, Czech Republic, October 28 to 31, 2012. Submitted.
15. Soliman, A. M. and Nehdi, M. L. (2011), “Environmentally friendly self-accelerated concrete,” Proceedings of 2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted.
16. Soliman, A. M. and Nehdi, M. L. (2011), “Early-age behaviour of shrinkage reducing admixture and wollastonite microfibers in UHPC,” Proceedings of
380
2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted.
17. Soliman, A. M. and Nehdi, M. L. (2011), “Effect of exposure conditions on early-age shrinkage of UHPC thin elements,” Proceedings of 2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted.
18. J. Camiletti, Soliman, A.M. and M.L. Nehdi (2010). “Early age behaviour of UHPC incorporating micro-limestone,” Proceedings of 2nd International Engineering Mechanics and Materials Specialty Conference, Ottawa, Ontario, Canada. Submitted (Co-supervising Work).
HONOURS AND AWARDS
• 2010-2011: Graduate Research Thesis Award: University of Western Ontario, London, Ontario, Canada
• 2009-2010: Graduate Research Thesis Award: University of Western Ontario,
London, Ontario, Canada
• 2009-2010: Graduate scholarship in structure Engineering Research award, The University of Western Ontario, London, Ontario, Canada
• 2007-Present: Western Engineering Graduate Entrance Scholarship (WEGES),
The University of Western Ontario, London, Ontario, Canada
• 1997-2002: The Government Award of Excellence in Undergraduate Studies,
Ain Shams University, Cairo, Egypt.
TEACHING EXPERIENCES
• 2008-2010: Instructing a number of lectures in undergraduate courses: Mechanics of Materials [Second year course] & Materials for Civil Engineering [Third year course], The University of Western Ontario
381
SUPERVISORY EXPERIENCES
• 2009- 2011: Co-supervising a Master student working on “The application of nano-limestone in concrete” at The University of Western Ontario [Publication is mentioned above].
• 2008-2009: Supervising three summer students working at the Concrete Laboratory in the Faculty of Engineering at The University of Western Ontario.
OTHER RELATED EXPERIENCE
• 2004-2007 Designer, (Part Time) Engineering Consulting Center, Ain Shams University, Cairo, Egypt. Duties: Design concrete and steel structures, analysis for several normal and medium-rise buildings (up to 12 story), modeling and preparing structural drawings. Studying different problems in the area of industry and engineering and create solution methods.
• 2002-2007 Laboratory and quality control engineer, Materials laboratory, Ain Shams University, Cairo, Egypt Duties: supervision and conducting experiments, site visiting, test results analysis and writing technical reports.
• 2007-Present Teaching Assistant and Research Assistant, The University Of Western Ontario, London, Ontario, Canada • 2002-2007 Teaching Assistant and Research Assistant,
Ain Shams University, Cairo, Egypt Courses: Properties & Testing of Materials (1&2), New Construction Materials, Special Concrete Types, Concrete Durability, Repair & Strengthening of Structures.
• 2006 Research Assistant, (Part-Time)
National Research Center, Cairo, Egypt Duties: Studying the behavior of fiber reinforced concrete exposed to fire.
ELECTED POSITIONS
• 2009- 2011: Councilor for the Faculty of Engineering for the Society of Graduate Students (SOGS)-4000 members, The University of Western Ontario.
382
• 2010- 2011: Councilor for the Graduate Engineering Society-1500 members,
The University of Western Ontario.
VOLUNTEER WORK
• 2009-2010: Concrete Canoe 2009: provide information and advice to team members on different concrete practices.
• 2008-2009: Concrete lab assistant: casting concrete samples to be used for demonstration purposes for undergraduate and postgraduate students.
• 2009-Present: Steward for Civil and Environmental Engineering Department at the Graduate Teaching Assistant Union, The University of Western Ontario.
• 2009-Present: Social Activities Coordinator at the Egyptian Student Association in North America (ESANA), London, Ontario, Canada.
MEMBERSHIPS:
• (CSCE) The Canadian Society for Civil Engineering. • (ACI) The American Concrete Institute. • (ASCE) American Society of Civil Engineers. • (ESE) Egyptian Syndicate of Engineers.